2013-08-06 13:38:00 +00:00
|
|
|
"""
|
|
|
|
mbed SDK
|
2016-02-05 07:19:48 +00:00
|
|
|
Copyright (c) 2011-2016 ARM Limited
|
2013-08-06 13:38:00 +00:00
|
|
|
|
|
|
|
Licensed under the Apache License, Version 2.0 (the "License");
|
|
|
|
you may not use this file except in compliance with the License.
|
|
|
|
You may obtain a copy of the License at
|
|
|
|
|
2014-04-02 09:16:53 +00:00
|
|
|
http://www.apache.org/licenses/LICENSE-2.0
|
2013-08-06 13:38:00 +00:00
|
|
|
|
|
|
|
Unless required by applicable law or agreed to in writing, software
|
|
|
|
distributed under the License is distributed on an "AS IS" BASIS,
|
|
|
|
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
|
|
See the License for the specific language governing permissions and
|
|
|
|
limitations under the License.
|
|
|
|
"""
|
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
import os
|
|
|
|
import binascii
|
|
|
|
import struct
|
|
|
|
import shutil
|
|
|
|
import inspect
|
|
|
|
import sys
|
|
|
|
from tools.patch import patch
|
|
|
|
from tools.paths import TOOLS_BOOTLOADERS
|
|
|
|
from tools.utils import json_file_to_dict
|
|
|
|
|
2013-06-24 13:32:08 +00:00
|
|
|
CORE_LABELS = {
|
2016-02-05 07:19:48 +00:00
|
|
|
"ARM7TDMI-S": ["ARM7", "LIKE_CORTEX_ARM7"],
|
|
|
|
"Cortex-M0" : ["M0", "CORTEX_M", "LIKE_CORTEX_M0"],
|
|
|
|
"Cortex-M0+": ["M0P", "CORTEX_M", "LIKE_CORTEX_M0"],
|
|
|
|
"Cortex-M1" : ["M1", "CORTEX_M", "LIKE_CORTEX_M1"],
|
|
|
|
"Cortex-M3" : ["M3", "CORTEX_M", "LIKE_CORTEX_M3"],
|
|
|
|
"Cortex-M4" : ["M4", "CORTEX_M", "RTOS_M4_M7", "LIKE_CORTEX_M4"],
|
2016-07-08 15:16:12 +00:00
|
|
|
"Cortex-M4F" : ["M4", "CORTEX_M", "RTOS_M4_M7", "LIKE_CORTEX_M4"],
|
2016-02-05 07:19:48 +00:00
|
|
|
"Cortex-M7" : ["M7", "CORTEX_M", "RTOS_M4_M7", "LIKE_CORTEX_M7"],
|
2016-07-08 15:16:12 +00:00
|
|
|
"Cortex-M7F" : ["M7", "CORTEX_M", "RTOS_M4_M7", "LIKE_CORTEX_M7"],
|
|
|
|
"Cortex-M7FD" : ["M7", "CORTEX_M", "RTOS_M4_M7", "LIKE_CORTEX_M7"],
|
2016-02-05 07:19:48 +00:00
|
|
|
"Cortex-A9" : ["A9", "CORTEX_A", "LIKE_CORTEX_A9"]
|
2013-06-24 13:32:08 +00:00
|
|
|
}
|
|
|
|
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
# Generic Target class that reads and interprets the data in targets.json
|
|
|
|
|
|
|
|
class HookError(Exception):
|
2016-08-12 21:52:26 +00:00
|
|
|
""" A simple class that represents all the exceptions associated with
|
|
|
|
hooking
|
|
|
|
"""
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
pass
|
2014-05-29 13:42:03 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
CACHES = {}
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
def cached(func):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""A simple decorator used for automatically caching data returned by a
|
|
|
|
function
|
|
|
|
"""
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
def wrapper(*args, **kwargs):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""The wrapped function itself"""
|
|
|
|
if not CACHES.has_key((func.__name__, args)):
|
|
|
|
CACHES[(func.__name__, args)] = func(*args, **kwargs)
|
|
|
|
return CACHES[(func.__name__, args)]
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
return wrapper
|
2014-05-29 13:42:03 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
class Target(object):
|
|
|
|
"""An object to represent a Target (MCU/Board)"""
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
# Cumulative attributes can have values appended to them, so they
|
|
|
|
# need to be computed differently than regular attributes
|
2016-08-12 21:49:54 +00:00
|
|
|
cumulative_attributes = ['extra_labels', 'macros', 'device_has', 'features']
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
# List of targets that were added dynamically using "add_py_targets" (see
|
|
|
|
# below)
|
2016-07-19 10:14:42 +00:00
|
|
|
__py_targets = set()
|
|
|
|
|
2016-08-02 10:42:21 +00:00
|
|
|
# Default location of the 'targets.json' file
|
2016-08-12 21:52:26 +00:00
|
|
|
__targets_json_location_default = os.path.join(
|
|
|
|
os.path.dirname(os.path.abspath(__file__)), '..', 'hal', 'targets.json')
|
2016-08-02 10:42:21 +00:00
|
|
|
|
|
|
|
# Current/new location of the 'targets.json' file
|
|
|
|
__targets_json_location = None
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
|
|
|
|
@staticmethod
|
|
|
|
@cached
|
|
|
|
def get_json_target_data():
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Load the description of JSON target data"""
|
|
|
|
return json_file_to_dict(Target.__targets_json_location or
|
|
|
|
Target.__targets_json_location_default)
|
2016-07-12 09:58:53 +00:00
|
|
|
|
|
|
|
@staticmethod
|
2016-08-02 10:42:21 +00:00
|
|
|
def set_targets_json_location(location=None):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Set the location of the targets.json file"""
|
|
|
|
Target.__targets_json_location = (location or
|
|
|
|
Target.__targets_json_location_default)
|
2016-07-12 09:58:53 +00:00
|
|
|
# Invalidate caches, since the location of the JSON file changed
|
2016-08-12 21:52:26 +00:00
|
|
|
CACHES.clear()
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
|
|
|
|
@staticmethod
|
|
|
|
@cached
|
|
|
|
def get_module_data():
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Get the members of this module using Python's "inspect" module"""
|
|
|
|
return dict([(m[0], m[1]) for m in
|
|
|
|
inspect.getmembers(sys.modules[__name__])])
|
|
|
|
|
|
|
|
def __get_resolution_order(self, target_name, order, level=0):
|
|
|
|
""" Return the order in which target descriptions are searched for
|
|
|
|
attributes. This mimics the Python 2.2 method resolution order, which
|
|
|
|
is what the old targets.py module used. For more details, check
|
|
|
|
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
|
|
|
|
The resolution order contains (name, level) tuples, where "name" is the
|
|
|
|
name of the class and "level" is the level in the inheritance hierarchy
|
|
|
|
(the target itself is at level 0, its first parent at level 1, its
|
|
|
|
parent's parent at level 2 and so on)
|
|
|
|
"""
|
|
|
|
# the resolution order can't contain duplicate target names
|
2016-08-15 22:30:50 +00:00
|
|
|
if target_name not in [l[0] for l in order]:
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
order.append((target_name, level))
|
|
|
|
parents = self.get_json_target_data()[target_name].get("inherits", [])
|
2016-08-12 21:52:26 +00:00
|
|
|
for par in parents:
|
|
|
|
order = self.__get_resolution_order(par, order, level + 1)
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
return order
|
|
|
|
|
|
|
|
@staticmethod
|
|
|
|
def __add_paths_to_progen(data):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Modify the exporter specification ("progen") by changing all
|
|
|
|
"template" keys to full paths
|
|
|
|
"""
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
out = {}
|
2016-08-12 21:52:26 +00:00
|
|
|
for key, val in data.items():
|
|
|
|
if isinstance(val, dict):
|
|
|
|
out[key] = Target.__add_paths_to_progen(val)
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
elif key == "template":
|
2016-08-12 21:52:26 +00:00
|
|
|
out[key] = [os.path.join(os.path.dirname(__file__), 'export', v)
|
|
|
|
for v in val]
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
else:
|
2016-08-12 21:52:26 +00:00
|
|
|
out[key] = val
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
return out
|
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
def __getattr_cumulative(self, attrname):
|
|
|
|
"""Look for the attribute in the class and its parents, as defined by
|
|
|
|
the resolution order
|
|
|
|
"""
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
tdata = self.get_json_target_data()
|
2016-08-12 21:52:26 +00:00
|
|
|
# For a cumulative attribute, figure out when it was defined the
|
|
|
|
# last time (in attribute resolution order) then follow the "_add"
|
|
|
|
# and "_remove" data fields
|
|
|
|
for idx, target in enumerate(self.resolution_order):
|
|
|
|
# the attribute was defined at this level in the resolution
|
|
|
|
# order
|
|
|
|
if attrname in tdata[target[0]]:
|
|
|
|
def_idx = idx
|
|
|
|
break
|
|
|
|
else:
|
|
|
|
raise AttributeError("Attribute '%s' not found in target '%s'"
|
2016-08-15 22:30:50 +00:00
|
|
|
% (attrname, self.name))
|
2016-08-12 21:52:26 +00:00
|
|
|
# Get the starting value of the attribute
|
|
|
|
starting_value = (tdata[self.resolution_order[def_idx][0]][attrname]
|
2016-08-15 22:30:50 +00:00
|
|
|
or [])[:]
|
2016-08-12 21:52:26 +00:00
|
|
|
# Traverse the resolution list in high inheritance to low
|
|
|
|
# inheritance level, left to right order to figure out all the
|
|
|
|
# other classes that change the definition by adding or removing
|
|
|
|
# elements
|
|
|
|
for idx in xrange(self.resolution_order[def_idx][1] - 1, -1, -1):
|
|
|
|
same_level_targets = [tar[0] for tar in self.resolution_order
|
2016-08-15 22:30:50 +00:00
|
|
|
if tar[1] == idx]
|
2016-08-12 21:52:26 +00:00
|
|
|
for tar in same_level_targets:
|
|
|
|
data = tdata[tar]
|
|
|
|
# Do we have anything to add ?
|
|
|
|
if data.has_key(attrname + "_add"):
|
|
|
|
starting_value.extend(data[attrname + "_add"])
|
|
|
|
# Do we have anything to remove ?
|
|
|
|
if data.has_key(attrname + "_remove"):
|
|
|
|
# Macros can be defined either without a value (MACRO)
|
|
|
|
# or with a value (MACRO=10). When removing, we specify
|
|
|
|
# only the name of the macro, without the value. So we
|
|
|
|
# need to create a mapping between the macro name and
|
|
|
|
# its value. This will work for extra_labels and other
|
|
|
|
# type of arrays as well, since they fall into the
|
|
|
|
# "macros without a value" category (simple definitions
|
|
|
|
# without a value).
|
|
|
|
name_def_map = {}
|
|
|
|
for crtv in starting_value:
|
|
|
|
if crtv.find('=') != -1:
|
|
|
|
temp = crtv.split('=')
|
|
|
|
if len(temp) != 2:
|
|
|
|
raise ValueError(
|
|
|
|
"Invalid macro definition '%s'" % crtv)
|
|
|
|
name_def_map[temp[0]] = crtv
|
|
|
|
else:
|
|
|
|
name_def_map[crtv] = crtv
|
|
|
|
for element in data[attrname + "_remove"]:
|
2016-08-15 22:30:50 +00:00
|
|
|
if element not in name_def_map:
|
2016-08-12 21:52:26 +00:00
|
|
|
raise ValueError(
|
|
|
|
("Unable to remove '%s' in '%s.%s' since "
|
2016-08-15 22:30:50 +00:00
|
|
|
% (element, self.name, attrname)) +
|
2016-08-12 21:52:26 +00:00
|
|
|
"it doesn't exist")
|
|
|
|
starting_value.remove(name_def_map[element])
|
|
|
|
return starting_value
|
|
|
|
|
|
|
|
def __getattr_helper(self, attrname):
|
|
|
|
"""Compute the value of a given target attribute"""
|
|
|
|
if attrname in self.cumulative_attributes:
|
|
|
|
return self.__getattr_cumulative(attrname)
|
|
|
|
else:
|
|
|
|
tdata = self.get_json_target_data()
|
|
|
|
starting_value = None
|
|
|
|
for target in self.resolution_order:
|
|
|
|
data = tdata[target[0]]
|
|
|
|
if data.has_key(attrname):
|
|
|
|
starting_value = data[attrname]
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
break
|
2016-08-12 21:52:26 +00:00
|
|
|
else: # Attribute not found
|
|
|
|
raise AttributeError(
|
|
|
|
"Attribute '%s' not found in target '%s'"
|
|
|
|
% (attrname, self.name))
|
|
|
|
# 'progen' needs the full path to the template (the path in JSON is
|
|
|
|
# relative to tools/export)
|
|
|
|
if attrname == "progen":
|
|
|
|
return self.__add_paths_to_progen(starting_value)
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
else:
|
2016-08-12 21:52:26 +00:00
|
|
|
return starting_value
|
|
|
|
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
def __getattr__(self, attrname):
|
2016-08-12 21:52:26 +00:00
|
|
|
""" Return the value of an attribute. This function only computes the
|
|
|
|
attribute's value once, then adds it to the instance attributes (in
|
|
|
|
__dict__), so the next time it is returned directly
|
|
|
|
"""
|
|
|
|
result = self.__getattr_helper(attrname)
|
|
|
|
self.__dict__[attrname] = result
|
|
|
|
return result
|
2016-06-09 21:05:35 +00:00
|
|
|
|
|
|
|
@staticmethod
|
|
|
|
def add_py_targets(new_targets):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Add one or more new target(s) represented as a Python dictionary
|
|
|
|
in 'new_targets'. It is an error to add a target with a name that
|
|
|
|
already exists.
|
|
|
|
"""
|
2016-06-09 21:05:35 +00:00
|
|
|
crt_data = Target.get_json_target_data()
|
2016-08-12 21:52:26 +00:00
|
|
|
for target_key, target_value in new_targets.items():
|
|
|
|
if crt_data.has_key(target_key):
|
|
|
|
raise Exception(
|
|
|
|
"Attempt to add target '%s' that already exists"
|
|
|
|
% target_key)
|
2016-07-12 11:42:21 +00:00
|
|
|
# Add target data to the internal target dictionary
|
2016-08-12 21:52:26 +00:00
|
|
|
crt_data[target_key] = target_value
|
2016-07-12 11:42:21 +00:00
|
|
|
# Create the new target and add it to the relevant data structures
|
2016-08-12 21:52:26 +00:00
|
|
|
new_target = Target(target_key)
|
2016-07-12 11:42:21 +00:00
|
|
|
TARGETS.append(new_target)
|
2016-08-12 21:52:26 +00:00
|
|
|
TARGET_MAP[target_key] = new_target
|
|
|
|
TARGET_NAMES.append(target_key)
|
2016-06-09 21:05:35 +00:00
|
|
|
|
|
|
|
@staticmethod
|
2016-07-12 09:54:44 +00:00
|
|
|
@cached
|
2016-08-12 21:52:26 +00:00
|
|
|
def get_target(target_name):
|
|
|
|
""" Return the target instance starting from the target name """
|
|
|
|
return Target(target_name)
|
Target K64F addition.
Squashed commit of the following:
commit db3c9f7682083abeb291e01df31e67e4c50845b3
Author: 0xc0170 <c0170@rocketmail.com>
Date: Wed Apr 2 09:52:00 2014 +0200
K64F - KSDK - Warnings fixes
commit a639a5cdff889c13509c954b0a34ebac861c1361
Merge: 67a2c2a f3de345
Author: 0xc0170 <c0170@rocketmail.com>
Date: Tue Apr 1 12:48:35 2014 +0200
Merge branch latest 'master' into dev_target_k64f
Conflicts:
libraries/rtos/rtx/RTX_Conf_CM.c
workspace_tools/build_api.py
commit 67a2c2aeb976f264db52ea10d18fea9de0d7685f
Author: 0xc0170 <c0170@rocketmail.com>
Date: Sun Mar 30 13:19:51 2014 +0200
K64F - PinName for buttons (SW2, SW3)
commit 957573e2cd42d5c73ed99477abb98c8b883695b2
Author: 0xc0170 <c0170@rocketmail.com>
Date: Tue Mar 25 11:46:57 2014 +0100
K64F - pins addition to mbed HAL, uart - 4 instances, fix i2c instance which was not stored
commit 2347a6d03984e297190910a250f2771032ae6327
Author: sg- <sam.w.grove@gmail.com>
Date: Mon Mar 24 15:20:51 2014 -0500
Added wait to i2c stop
commit b7b4a9c72e8724087a44078c41a2cb33e4c8d5e3
Author: 0xc0170 <c0170@rocketmail.com>
Date: Mon Mar 24 19:28:16 2014 +0100
K64F - I2c - ack flag retreive correction (logic inverted in ksdk hal)
commit 46c875251263029e32463c3b48473b10496088d9
Author: sg- <sam.w.grove@gmail.com>
Date: Mon Mar 24 13:16:18 2014 -0500
Added I2C Pinnames
commit b71c7a0dfba7025662f9a9d176494ce4dc86273e
Author: 0xc0170 <c0170@rocketmail.com>
Date: Tue Mar 18 17:02:34 2014 +0100
K64F Arduino pinNames update
commit d41b0ee493263d1d80fcb72b3f0d4d788359c7c9
Merge: 9c0a982 e2574eb
Author: 0xc0170 <c0170@rocketmail.com>
Date: Tue Mar 18 14:57:57 2014 +0100
Merge remote-tracking branch 'upstream/master' into dev_target_k64f.
K64F gpio changed according to the latest mbed master.
Conflicts:
libraries/rtos/rtx/RTX_CM_lib.h
workspace_tools/export/uvision4.py
commit 9c0a9822ab14263fff5e3b6459b7c2b4a77ce30c
Author: 0xc0170 <c0170@rocketmail.com>
Date: Mon Mar 17 21:08:17 2014 +0100
K64F - sleep support
commit 5edcf3603d2e7b00eedbb377203a054b7a01e51d
Author: 0xc0170 <c0170@rocketmail.com>
Date: Sun Mar 16 18:19:55 2014 +0100
K64F - pullup/down corrections, LED1 - R, LED2 - G, LED3 - B
commit a2b3b53a1474d32779654765cd1ce2ba2c6b2186
Author: 0xc0170 <c0170@rocketmail.com>
Date: Thu Mar 13 20:55:39 2014 +0100
K64F - SPI - delays are set, pin definition for tests
commit 1f3b3abe16f4afaaf1b75cb4bf3e3a9d5b6e50a7
Author: 0xc0170 <c0170@rocketmail.com>
Date: Tue Mar 11 21:26:00 2014 +0100
K64F - DAC update - tested with test a8
- internal reference set to VDDA
- PinName DAC0_OUT
commit 26d8cf47f8c0786b290ae659beb022901029b313
Author: 0xc0170 <c0170@rocketmail.com>
Date: Tue Mar 11 08:31:44 2014 +0100
KSDK - drivers layer removal, mbed HAL using only KSDK HAL
- ADC corrections with channels, and clock configuration
commit 67ebe14f5c88725033ea0fb135d94b6bf9a00fdb
Author: 0xc0170 <c0170@rocketmail.com>
Date: Mon Mar 10 12:46:08 2014 +0100
K20 copy files removed, targets.py - vertical alignment for K64F definition
commit be5c562d513c808e5bd425195a4fb1c71f47a57e
Merge: 696a713 fe0aca9
Author: Emilio Monti <emilmont@gmail.com>
Date: Mon Mar 10 11:14:55 2014 +0000
Merge branch 'rtos_support' into dev_target_k64f
commit 696a713751194b4762f1cdf6c17c0786decd7808
Author: 0xc0170 <c0170@rocketmail.com>
Date: Mon Mar 10 12:05:30 2014 +0100
[FIX] K64F - adc, sgtl driver updates
commit fe0aca9940bbdd5ee70a1a7341a0a2ad2abf912b
Author: Emilio Monti <emilmont@gmail.com>
Date: Mon Mar 10 11:04:16 2014 +0000
Add RTOS support for K64F
commit 5c3edcbca6dbcce628c7cde51ac94a6fc6278ba5
Author: 0xc0170 <c0170@rocketmail.com>
Date: Sun Mar 9 20:43:38 2014 +0100
K64F - uvision templates update
- uvision 5.10 supports K64F
commit 33f18d11d0eadb9243f1be0ae96c5f82e2913d48
Merge: 26f7587 74e9b2e
Author: 0xc0170 <c0170@rocketmail.com>
Date: Sat Mar 8 10:34:25 2014 +0100
Update branch from mbed master
- merge branch 'master' into dev_target_k64f
Conflicts:
libraries/USBDevice/USBDevice/USBEndpoints.h
libraries/USBDevice/USBDevice/USBHAL_KL25Z.cpp
workspace_tools/export/uvision4.py
workspace_tools/targets.py
commit 26f75872b19a1db2a3abb34c6e773bac56acb32f
Author: 0xc0170 <c0170@rocketmail.com>
Date: Thu Mar 6 22:15:53 2014 +0100
K64F - USBDevice - MPU disable in init
commit e54d6bbaa68827bd63058fbf2428e289d15ac1f7
Author: 0xc0170 <c0170@rocketmail.com>
Date: Wed Feb 26 21:06:58 2014 +0100
K64F - clock setup 4 (usb clock enable)
commit c4165732b9520cb31ae3d649d50c353d09dc9932
Author: 0xc0170 <c0170@rocketmail.com>
Date: Wed Feb 26 20:01:47 2014 +0100
K64F - USBDevice addition
commit 9fcdbb8562b1415561b04e902fcdbb4724add5af
Author: 0xc0170 <c0170@rocketmail.com>
Date: Mon Feb 24 19:11:48 2014 +0100
K64F SPI HAL - initial version
commit 8093df58fa7d17fcb5ad04872c958d5254ee1d8a
Author: 0xc0170 <c0170@rocketmail.com>
Date: Sat Feb 22 13:14:44 2014 +0100
K64F - RTC and I2C implementation (using KPSDK HAL)
- instance correction in objects and peripherals names headers
commit 7ef3fbda605c2bd53a86f37d0676c0393b2e2949
Author: 0xc0170 <c0170@rocketmail.com>
Date: Fri Feb 14 09:45:27 2014 +0100
mbed RTC HAL for K64F
commit e40332fd2db8bf36b3e6cabac5729e013da40c28
Merge: e059f65 6bfcd87
Author: 0xc0170 <c0170@rocketmail.com>
Date: Thu Feb 13 14:20:20 2014 +0100
Merge branch 'master' into dev_target_k64f
Conflicts:
workspace_tools/build_api.py
workspace_tools/export/uvision4.py
workspace_tools/targets.py
workspace_tools/toolchains/__init__.py
commit e059f65fd09694418f9fa4f38da90954ab9decfe
Author: 0xc0170 <c0170@rocketmail.com>
Date: Wed Feb 5 21:35:49 2014 +0100
pwm mbed HAL
- using jusst ftm hal from KPSDK, not yet functional
commit b784278872b1d66ce2940f4988e0479971de8bc0
Author: 0xc0170 <c0170@rocketmail.com>
Date: Mon Feb 3 18:28:24 2014 +0100
uvision exporters for K64F, uticker - lptmr
- lptmr - no hal neiter driver, quick implementation using registers and internal clock
- exporters for K64F - using K60 1MB target, because K64F is not available in 4.7 yet
commit 7a030693e025c2bd456563f3e6f4456033d3f644
Author: Bogdan Marinescu <bogdan.marinescu@arm.com>
Date: Tue Jan 28 16:29:54 2014 +0200
Attempt to keep target's include directory structure
commit a2445b383224125abf4ee23bd17f1e685010a4a5
Author: 0xc0170 <c0170@rocketmail.com>
Date: Mon Jan 27 07:25:16 2014 +0100
Original KPSDK include directory structure for device
commit 9c07c58bb9cf5e9d8be4c3bec117ee87a5ea81c0
Author: 0xc0170 <c0170@rocketmail.com>
Date: Fri Jan 24 16:51:17 2014 +0000
K64F ADC - initial commit
- ADC using KPSDK driver
commit 88e03ef8c5855a57887bb36cddfa5ab1491d400c
Author: 0xc0170 <c0170@rocketmail.com>
Date: Fri Jan 24 12:18:14 2014 +0000
GPI IRQ
- nvic vectors number correction
- gpio irq HAL implementation
commit e83f1108ae9f779ce240d6cdfe23532bfa00a55e
Author: 0xc0170 <c0170@rocketmail.com>
Date: Fri Jan 24 10:06:44 2014 +0000
PORT HAL implementation
- using gpio hal port
commit 75c21943200c8240d1edaf0a67f84b9f3b43db7f
Author: 0xc0170 <c0170@rocketmail.com>
Date: Thu Jan 23 16:02:36 2014 +0000
Serial (only 8bit at the moment), using KPSDK HAL
commit 296e79def617f005918cd8e2aa574f2908a362ca
Author: 0xc0170 <c0170@rocketmail.com>
Date: Thu Jan 23 08:35:50 2014 +0000
Folder structure correction for K64F - KPSDK
- scripts reverted, only new macro is available
- K64F specific headers are in HAL in device folder
commit f236b1ffcb9c8b443ad8483bca8b0e564a63f004
Author: 0xc0170 <c0170@rocketmail.com>
Date: Wed Jan 22 16:07:30 2014 +0100
update KPSDK to RC1
- the update causes few more dependences which were reported. Will be removed later (usb, boards)
- pit timer - hal use , pit driver uses us period
commit f02c5353d4920e885f803ad235e5e82001e97b94
Author: 0xc0170 <c0170@rocketmail.com>
Date: Tue Jan 21 09:45:55 2014 +0100
KPSDK In/out declaration removal
commit 8c88e098b4dc4901753309f1e6db4adb8aca4384
Author: 0xc0170 <c0170@rocketmail.com>
Date: Tue Jan 21 09:12:41 2014 +0100
gpio_t struct only needs pinName
- gpio_init creates init objects on stack
commit 6b96d4ea2c5a6a2cb13571d740ffb679a62f8f3d
Author: 0xc0170 <c0170@rocketmail.com>
Date: Mon Jan 20 19:59:03 2014 +0100
us ticker - pit implementation (not functional yet)
- pit driver in KPSDK - added sdk prefix to needed functions
commit 098e60a3846abcd4c9c00bd199b01d4b1899807f
Author: 0xc0170 <c0170@rocketmail.com>
Date: Mon Jan 20 13:01:58 2014 +0100
GPIO HAL - gpio_set implementation
commit 2bfebbfc75dcd08c20297ba42dc0cc82e5381a40
Author: 0xc0170 <c0170@rocketmail.com>
Date: Sun Jan 19 20:46:55 2014 +0100
GPIO KPSDK changes
- gpio driver - sdk prefix, no lookuptable, input/output declaration, refactoring, set MUX to GPIO
- gpio api in mbed implementation, tested on blue led
commit d083733c485fbdd79ed9ce87100df9fee82294a7
Author: 0xc0170 <c0170@rocketmail.com>
Date: Sat Jan 18 17:14:09 2014 +0100
Update folder structure for KPSDK
- drivers addition
- usb (needed by drivers - needs to be fixed)
- utilities
- hal folder
- drivers/flash removed (needs to be fixed)
- usb host removed (needs to be fixed)
commit 9abcf3d94a2cc849cd6e586c1bad650b6a340a0c
Author: 0xc0170 <c0170@rocketmail.com>
Date: Thu Jan 16 11:06:16 2014 +0100
Initial commit for K64F
- KPSDK addition
- CMSIS + HAL for K64F
- HAL is not yet implemented
- scripts - target contain macros, ignore folders, cmsis copy folders
2014-04-02 12:39:01 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
def __init__(self, target_name):
|
|
|
|
self.name = target_name
|
2014-07-02 09:56:24 +00:00
|
|
|
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
# Compute resolution order once (it will be used later in __getattr__)
|
|
|
|
self.resolution_order = self.__get_resolution_order(self.name, [])
|
2016-08-12 21:52:26 +00:00
|
|
|
# Create also a list with only the names of the targets in the
|
|
|
|
# resolution order
|
|
|
|
self.resolution_order_names = [target[0] for target
|
|
|
|
in self.resolution_order]
|
2014-09-24 09:47:55 +00:00
|
|
|
|
2016-06-27 21:01:52 +00:00
|
|
|
@property
|
2013-07-25 16:55:52 +00:00
|
|
|
def program_cycle_s(self):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Special override for program_cycle_s as it's default value depends
|
|
|
|
upon is_disk_virtual
|
|
|
|
"""
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
try:
|
|
|
|
return self.__getattr__("program_cycle_s")
|
|
|
|
except AttributeError:
|
|
|
|
return 4 if self.is_disk_virtual else 1.5
|
2014-05-29 13:42:03 +00:00
|
|
|
|
2013-06-24 13:32:08 +00:00
|
|
|
def get_labels(self):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Get all possible labels for this target"""
|
2016-06-09 21:05:35 +00:00
|
|
|
labels = [self.name] + CORE_LABELS[self.core] + self.extra_labels
|
2016-08-12 21:52:26 +00:00
|
|
|
# Automatically define UVISOR_UNSUPPORTED if the target doesn't
|
|
|
|
# specifically define UVISOR_SUPPORTED
|
2016-08-15 22:30:50 +00:00
|
|
|
if "UVISOR_SUPPORTED" not in labels:
|
2016-06-09 21:05:35 +00:00
|
|
|
labels.append("UVISOR_UNSUPPORTED")
|
|
|
|
return labels
|
2014-05-29 13:42:03 +00:00
|
|
|
|
2013-08-30 09:19:08 +00:00
|
|
|
def init_hooks(self, hook, toolchain_name):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Initialize the post-build hooks for a toolchain. For now, this
|
|
|
|
function only allows "post binary" hooks (hooks that are executed
|
|
|
|
after the binary image is extracted from the executable file)
|
|
|
|
"""
|
|
|
|
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
# If there's no hook, simply return
|
|
|
|
try:
|
|
|
|
hook_data = self.post_binary_hook
|
|
|
|
except AttributeError:
|
|
|
|
return
|
2016-08-12 21:52:26 +00:00
|
|
|
# A hook was found. The hook's name is in the format
|
|
|
|
# "classname.functionname"
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
temp = hook_data["function"].split(".")
|
|
|
|
if len(temp) != 2:
|
2016-08-12 21:52:26 +00:00
|
|
|
raise HookError(
|
|
|
|
("Invalid format for hook '%s' in target '%s'"
|
|
|
|
% (hook_data["function"], self.name)) +
|
|
|
|
" (must be 'class_name.function_name')")
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
class_name, function_name = temp[0], temp[1]
|
2016-08-12 21:52:26 +00:00
|
|
|
# "class_name" must refer to a class in this file, so check if the
|
|
|
|
# class exists
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
mdata = self.get_module_data()
|
2016-08-12 21:52:26 +00:00
|
|
|
if not mdata.has_key(class_name) or \
|
|
|
|
not inspect.isclass(mdata[class_name]):
|
|
|
|
raise HookError(
|
|
|
|
("Class '%s' required by '%s' in target '%s'"
|
|
|
|
% (class_name, hook_data["function"], self.name)) +
|
|
|
|
" not found in targets.py")
|
|
|
|
# "function_name" must refer to a static function inside class
|
|
|
|
# "class_name"
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
cls = mdata[class_name]
|
2016-08-12 21:52:26 +00:00
|
|
|
if (not hasattr(cls, function_name)) or \
|
|
|
|
(not inspect.isfunction(getattr(cls, function_name))):
|
|
|
|
raise HookError(
|
|
|
|
("Static function '%s' " % function_name) +
|
|
|
|
("required by '%s' " % hook_data["function"]) +
|
|
|
|
("in target '%s' " % self.name) +
|
|
|
|
("not found in class '%s'" % class_name))
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
# Check if the hook specification also has target restrictions
|
|
|
|
toolchain_restrictions = hook_data.get("toolchains", [])
|
2016-08-12 21:52:26 +00:00
|
|
|
if toolchain_restrictions and \
|
|
|
|
(toolchain_name not in toolchain_restrictions):
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
return
|
|
|
|
# Finally, hook the requested function
|
|
|
|
hook.hook_add_binary("post", getattr(cls, function_name))
|
2014-08-12 13:59:50 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
################################################################################
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
# Target specific code goes in this section
|
2016-08-12 21:52:26 +00:00
|
|
|
# This code can be invoked from the target description using the
|
|
|
|
# "post_binary_hook" key
|
2014-08-12 13:59:50 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
class LPCTargetCode(object):
|
|
|
|
"""General LPC Target patching code"""
|
2014-08-12 13:59:50 +00:00
|
|
|
@staticmethod
|
|
|
|
def lpc_patch(t_self, resources, elf, binf):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Patch an elf file"""
|
2014-08-12 13:59:50 +00:00
|
|
|
t_self.debug("LPC Patch: %s" % os.path.split(binf)[1])
|
|
|
|
patch(binf)
|
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
class LPC4088Code(object):
|
|
|
|
"""Code specific to the LPC4088"""
|
2013-08-30 09:19:08 +00:00
|
|
|
@staticmethod
|
2014-02-11 10:50:16 +00:00
|
|
|
def binary_hook(t_self, resources, elf, binf):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Hook to be run after an elf file is built"""
|
2013-08-30 09:19:08 +00:00
|
|
|
if not os.path.isdir(binf):
|
|
|
|
# Regular binary file, nothing to do
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
LPCTargetCode.lpc_patch(t_self, resources, elf, binf)
|
2013-08-30 09:19:08 +00:00
|
|
|
return
|
|
|
|
outbin = open(binf + ".temp", "wb")
|
|
|
|
partf = open(os.path.join(binf, "ER_IROM1"), "rb")
|
|
|
|
# Pad the fist part (internal flash) with 0xFF to 512k
|
|
|
|
data = partf.read()
|
|
|
|
outbin.write(data)
|
|
|
|
outbin.write('\xFF' * (512*1024 - len(data)))
|
|
|
|
partf.close()
|
2016-08-15 22:30:50 +00:00
|
|
|
# Read and append the second part (external flash) in chunks of fixed
|
|
|
|
# size
|
2013-08-30 09:19:08 +00:00
|
|
|
chunksize = 128 * 1024
|
|
|
|
partf = open(os.path.join(binf, "ER_IROM2"), "rb")
|
|
|
|
while True:
|
|
|
|
data = partf.read(chunksize)
|
|
|
|
outbin.write(data)
|
|
|
|
if len(data) < chunksize:
|
|
|
|
break
|
|
|
|
partf.close()
|
|
|
|
outbin.close()
|
|
|
|
# Remove the directory with the binary parts and rename the temporary
|
|
|
|
# file to 'binf'
|
|
|
|
shutil.rmtree(binf, True)
|
|
|
|
os.rename(binf + '.temp', binf)
|
|
|
|
t_self.debug("Generated custom binary file (internal flash + SPIFI)")
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
LPCTargetCode.lpc_patch(t_self, resources, elf, binf)
|
2015-02-07 02:14:26 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
class TEENSY3_1Code(object):
|
|
|
|
"""Hooks for the TEENSY3.1"""
|
2015-01-15 19:11:40 +00:00
|
|
|
@staticmethod
|
|
|
|
def binary_hook(t_self, resources, elf, binf):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Hook that is run after elf is generated"""
|
2015-01-15 19:11:40 +00:00
|
|
|
from intelhex import IntelHex
|
|
|
|
binh = IntelHex()
|
2016-08-12 21:52:26 +00:00
|
|
|
binh.loadbin(binf, offset=0)
|
2015-02-07 02:14:26 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
with open(binf.replace(".bin", ".hex"), "w") as file_desc:
|
|
|
|
binh.tofile(file_desc, format='hex')
|
2014-11-04 13:00:42 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
class MTSCode(object):
|
|
|
|
"""Generic MTS code"""
|
2015-02-13 19:25:31 +00:00
|
|
|
@staticmethod
|
2016-08-15 22:30:50 +00:00
|
|
|
def _combine_bins_helper(target_name, binf):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""combine bins with the bootloader for a particular target"""
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
loader = os.path.join(TOOLS_BOOTLOADERS, target_name, "bootloader.bin")
|
2015-02-13 19:25:31 +00:00
|
|
|
target = binf + ".tmp"
|
|
|
|
if not os.path.exists(loader):
|
|
|
|
print "Can't find bootloader binary: " + loader
|
|
|
|
return
|
|
|
|
outbin = open(target, 'w+b')
|
|
|
|
part = open(loader, 'rb')
|
|
|
|
data = part.read()
|
|
|
|
outbin.write(data)
|
|
|
|
outbin.write('\xFF' * (64*1024 - len(data)))
|
|
|
|
part.close()
|
|
|
|
part = open(binf, 'rb')
|
|
|
|
data = part.read()
|
|
|
|
outbin.write(data)
|
|
|
|
part.close()
|
|
|
|
outbin.seek(0, 0)
|
|
|
|
data = outbin.read()
|
2015-04-01 20:29:49 +00:00
|
|
|
outbin.seek(0, 1)
|
2015-02-13 19:25:31 +00:00
|
|
|
crc = struct.pack('<I', binascii.crc32(data) & 0xFFFFFFFF)
|
|
|
|
outbin.write(crc)
|
|
|
|
outbin.close()
|
|
|
|
os.remove(binf)
|
|
|
|
os.rename(target, binf)
|
|
|
|
|
2015-01-13 21:13:45 +00:00
|
|
|
@staticmethod
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
def combine_bins_mts_dot(t_self, resources, elf, binf):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""A hook for the MTS MDOT"""
|
2016-08-15 22:30:50 +00:00
|
|
|
MTSCode._combine_bins_helper("MTS_MDOT_F411RE", binf)
|
2014-07-02 09:56:24 +00:00
|
|
|
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
@staticmethod
|
|
|
|
def combine_bins_mts_dragonfly(t_self, resources, elf, binf):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""A hoof for the MTS Dragonfly"""
|
2016-08-15 22:30:50 +00:00
|
|
|
MTSCode._combine_bins_helper("MTS_DRAGONFLY_F411RE", binf)
|
2014-05-29 13:42:03 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
class MCU_NRF51Code(object):
|
|
|
|
"""NRF51 Hooks"""
|
2014-01-14 17:15:31 +00:00
|
|
|
@staticmethod
|
2016-08-15 22:30:50 +00:00
|
|
|
def binary_hook(t_self, resources, _, binf):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Hook that merges the soft device with the bin file"""
|
2015-07-23 12:28:52 +00:00
|
|
|
# Scan to find the actual paths of soft device
|
2015-03-06 16:54:14 +00:00
|
|
|
sdf = None
|
2016-08-12 21:52:26 +00:00
|
|
|
for softdevice_and_offset_entry\
|
|
|
|
in t_self.target.EXPECTED_SOFTDEVICES_WITH_OFFSETS:
|
2015-07-22 12:08:46 +00:00
|
|
|
for hexf in resources.hex_files:
|
2016-08-12 21:52:26 +00:00
|
|
|
if hexf.find(softdevice_and_offset_entry['name']) != -1:
|
|
|
|
t_self.debug("SoftDevice file found %s."
|
|
|
|
% softdevice_and_offset_entry['name'])
|
2015-07-23 12:28:52 +00:00
|
|
|
sdf = hexf
|
2015-03-06 16:54:14 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
if sdf is not None:
|
|
|
|
break
|
|
|
|
if sdf is not None:
|
|
|
|
break
|
2015-07-22 12:08:46 +00:00
|
|
|
|
2015-03-06 16:54:14 +00:00
|
|
|
if sdf is None:
|
2014-02-07 10:08:58 +00:00
|
|
|
t_self.debug("Hex file not found. Aborting.")
|
2014-01-14 17:15:31 +00:00
|
|
|
return
|
2016-02-26 00:29:38 +00:00
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
# Look for bootloader file that matches this soft device or bootloader
|
|
|
|
# override image
|
2015-07-23 12:28:52 +00:00
|
|
|
blf = None
|
|
|
|
if t_self.target.MERGE_BOOTLOADER is True:
|
|
|
|
for hexf in resources.hex_files:
|
|
|
|
if hexf.find(t_self.target.OVERRIDE_BOOTLOADER_FILENAME) != -1:
|
2016-08-12 21:52:26 +00:00
|
|
|
t_self.debug("Bootloader file found %s."
|
|
|
|
% t_self.target.OVERRIDE_BOOTLOADER_FILENAME)
|
2015-07-23 12:28:52 +00:00
|
|
|
blf = hexf
|
|
|
|
break
|
2016-08-12 21:52:26 +00:00
|
|
|
elif hexf.find(softdevice_and_offset_entry['boot']) != -1:
|
|
|
|
t_self.debug("Bootloader file found %s."
|
|
|
|
% softdevice_and_offset_entry['boot'])
|
2015-07-23 12:28:52 +00:00
|
|
|
blf = hexf
|
|
|
|
break
|
2014-05-29 13:42:03 +00:00
|
|
|
|
2014-02-21 15:05:21 +00:00
|
|
|
# Merge user code with softdevice
|
2014-01-14 17:15:31 +00:00
|
|
|
from intelhex import IntelHex
|
|
|
|
binh = IntelHex()
|
2016-08-12 21:52:26 +00:00
|
|
|
binh.loadbin(binf, offset=softdevice_and_offset_entry['offset'])
|
2014-05-29 13:42:03 +00:00
|
|
|
|
2014-11-19 12:45:36 +00:00
|
|
|
if t_self.target.MERGE_SOFT_DEVICE is True:
|
2016-08-12 21:52:26 +00:00
|
|
|
t_self.debug("Merge SoftDevice file %s"
|
|
|
|
% softdevice_and_offset_entry['name'])
|
2015-03-06 16:54:14 +00:00
|
|
|
sdh = IntelHex(sdf)
|
2014-11-19 12:45:36 +00:00
|
|
|
binh.merge(sdh)
|
2014-05-29 13:42:03 +00:00
|
|
|
|
2015-03-06 16:54:14 +00:00
|
|
|
if t_self.target.MERGE_BOOTLOADER is True and blf is not None:
|
2015-07-22 12:08:46 +00:00
|
|
|
t_self.debug("Merge BootLoader file %s" % blf)
|
2015-03-06 16:54:14 +00:00
|
|
|
blh = IntelHex(blf)
|
|
|
|
binh.merge(blh)
|
|
|
|
|
2016-08-12 21:52:26 +00:00
|
|
|
with open(binf.replace(".bin", ".hex"), "w") as fileout:
|
|
|
|
binh.tofile(fileout, format='hex')
|
2014-03-18 10:16:16 +00:00
|
|
|
|
2016-08-16 11:51:48 +00:00
|
|
|
class NCS36510TargetCode:
|
|
|
|
@staticmethod
|
|
|
|
def ncs36510_addfib(t_self, resources, elf, binf):
|
|
|
|
from tools.add_fib import add_fib_at_start
|
|
|
|
print("binf ", binf)
|
|
|
|
add_fib_at_start(binf[:-4])
|
2016-08-12 21:52:26 +00:00
|
|
|
################################################################################
|
2016-02-26 00:29:38 +00:00
|
|
|
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
# Instantiate all public targets
|
2016-08-12 21:52:26 +00:00
|
|
|
TARGETS = [Target.get_target(name) for name, value
|
|
|
|
in Target.get_json_target_data().items()
|
|
|
|
if value.get("public", True)]
|
2013-04-18 14:43:29 +00:00
|
|
|
|
|
|
|
# Map each target name to its unique instance
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
TARGET_MAP = dict([(t.name, t) for t in TARGETS])
|
2013-04-18 14:43:29 +00:00
|
|
|
|
|
|
|
TARGET_NAMES = TARGET_MAP.keys()
|
2013-11-18 18:24:51 +00:00
|
|
|
|
|
|
|
# Some targets with different name have the same exporters
|
Moved target definitions to JSON format
(long commit message ahead. Sorry about that, it can't be helped.)
This commit changs targets definition from Python to JSON format, as
part of the configuration mechanism implementation. There is a new file
under workspace_tools/ called "targets.json" which contains the target
definitions. "targets.py" remains, but becomes a wrapper on top of
"targets.json", with the same interface as before. This has the
advantage of not requiring code changes outside "targets.py".
Most of the JSON definitions of targets were automatically generated by a
script (available upon request since it doesn't make a lot of sense to
include it here), only those targets that had more than one parent in
the Python implementation were converted by hand. The target definitions
should be pretty self-explanatory. A number of things are different in
the JSON implementation (this is just a summary, check docs/mbed_targets.md
(also part of this PR) for a more complete description):
- "program_cycle_s" is now a value (as opposed to a function in the
Python implementation), since it only returned a number in all the
Python target implementations. The main definition that actually contains
some code (in class "Target") remains in target.py
- array values in "macros" and "extra_labels" can be modified
dynamically. Values can be added using "macros_add" and
"extra_labels_add" or removed using "macros_remove" and
"extra_labels_remove". This mechanism is available for all attributes
with a list type, but it's currently enabled only for "macros" and
"extra_labels" to keep things simple.
- "init_hooks"/"binary_hook" are now implemented in terms of a single
JSON key valled "post_binary_hook". The corresponding code is also in
"targets.py", under the various TargetCode classes (see for example
LPC4088Code in targets.py).
Just like in the Python implementation, a target can inherit from zero,
one or more targets. The resolution order for the target's attributes
follows the one used by the Python code (I used
http://makina-corpus.com/blog/metier/2014/python-tutorial-understanding-python-mro-class-search-path
as a reference for the implementation of resolution order).
This is obviously a very dangerous commit, since it affects all targets.
I tested compilation for a number of targets (K64F, LPC1768, NRF51822)
but there's definitely a lot more to be done in terms of testing.
I also tried to test in a different way: I wrote a script that imports the
old (Python) and the new (JSON) implementations and verifies that the
attributes in the old implementations exist and have the same values
in the new implementations (it also verifies that the attribute
resolution order is the same in the two implementations). If you're
interested, the script is here:
https://gist.github.com/bogdanm/c9d8cf34214109a4b9079befed6b3c0c
And the results of running the script are below (note that the script
outputs only the target names that were found to be problematic):
NRF51_MICROBIT_BOOT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_BOOT', 'MCU_NRF51_16K_BOOT_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BOOT_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_BOOT', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BOOT', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_BOOT', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT:
Resolution order is different in old and new
old: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT', 'MCU_NRF51_16K_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NRF51_16K_S110', 'MCU_NORDIC_16K', 'MCU_NRF51_16K']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_NRF_LFCLK_RC']
NRF51_MICROBIT_OTA:
Resolution order is different in old and new
old: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target', 'MCU_NRF51_S110']
new: ['NRF51_MICROBIT_OTA', 'MCU_NRF51_16K_OTA_S110', 'MCU_NRF51_S110', 'MCU_NRF51_16K_OTA_BASE', 'MCU_NRF51_16K_BASE', 'MCU_NRF51', 'Target']
'extra_labels' has different values in old and new
old: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_OTA', 'MCU_NRF51_16K_S110', 'NRF51_MICROBIT']
new: ['NORDIC', 'MCU_NRF51', 'MCU_NRF51822', 'MCU_NORDIC_16K', 'MCU_NRF51_16K', 'MCU_NRF51_16K_S110', 'MCU_NRF51_16K_OTA', 'NRF51_MICROBIT']
'macros' has different values in old and new
old: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
new: ['NRF51', 'TARGET_NRF51822', 'TARGET_MCU_NORDIC_16K', 'TARGET_MCU_NRF51_16K', 'TARGET_MCU_NRF51_16K_S110', 'TARGET_MCU_NRF51_16K_OTA', 'TARGET_OTA_ENABLED', 'TARGET_NRF51_MICROBIT', 'TARGET_NRF_LFCLK_RC']
NOT OK: ['NRF51_MICROBIT', 'NRF51_MICROBIT_BOOT', 'NRF51_MICROBIT_OTA']
The reasons for the above output are subtle and related to the
extremely weird way in which we defined target data in the Python
implementation: we used both class attributes and instance attributes.
This can complicate resolution order quite a bit and those two levels
don't exist in JSON: there's only one attribute type (equivalent to
Python's instance attributes). To make that work, I had to change the
inheritance order of the above targets (that use multiple inheritance)
which in turn changed the order of some macros and extra_labels (and of
course the resolution order). No harm done: the values are the same,
only their ordering is different. I don't believe this causes any
problems for 'extra_labels' and 'macros'.
This method of testing has its limitations though; in particular, it
can't test the hooks. I'm opened to ideas about how to test this better,
but I think that we need to remember that this commit might break some
targets and keep an eye out for "weird errors" in the future.
2016-05-17 17:50:01 +00:00
|
|
|
EXPORT_MAP = {}
|
2014-11-19 13:05:51 +00:00
|
|
|
|
|
|
|
# Detection APIs
|
|
|
|
def get_target_detect_codes():
|
|
|
|
""" Returns dictionary mapping detect_code -> platform_name
|
|
|
|
"""
|
|
|
|
result = {}
|
|
|
|
for target in TARGETS:
|
|
|
|
for detect_code in target.detect_code:
|
|
|
|
result[detect_code] = target.name
|
|
|
|
return result
|
2016-07-12 09:58:53 +00:00
|
|
|
|
2016-08-02 10:42:21 +00:00
|
|
|
def set_targets_json_location(location=None):
|
2016-08-12 21:52:26 +00:00
|
|
|
"""Sets the location of the JSON file that contains the targets"""
|
2016-07-12 09:58:53 +00:00
|
|
|
# First instruct Target about the new location
|
|
|
|
Target.set_targets_json_location(location)
|
2016-08-12 21:52:26 +00:00
|
|
|
# Then re-initialize TARGETS, TARGET_MAP and TARGET_NAMES. The
|
|
|
|
# re-initialization does not create new variables, it keeps the old ones
|
|
|
|
# instead. This ensures compatibility with code that does
|
|
|
|
# "from tools.targets import TARGET_NAMES"
|
2016-08-15 22:30:50 +00:00
|
|
|
TARGETS[:] = [Target.get_target(target) for target, obj
|
2016-08-12 21:52:26 +00:00
|
|
|
in Target.get_json_target_data().items()
|
2016-08-15 22:30:50 +00:00
|
|
|
if obj.get("public", True)]
|
2016-07-12 09:58:53 +00:00
|
|
|
TARGET_MAP.clear()
|
2016-08-12 21:52:26 +00:00
|
|
|
TARGET_MAP.update(dict([(target.name, target) for target in TARGETS]))
|
2016-07-12 09:58:53 +00:00
|
|
|
TARGET_NAMES[:] = TARGET_MAP.keys()
|