Explicitly setting the directory at the beginning of the start-mycroft.sh
and stop-mycroft.sh scripts. The behaves better if invoked from a
different directory.
Previously, the location of the log files changed depending
on whether you were running on a Mark1/Picroft or under a
"Github" install. Now they are always under the same directory
at /var/log/mycroft-*.log
This also updates the CLI to pull from that location always.
Additionally:
* Removed the nonfunctional 'wifi' option from start-mycroft.sh
* Made the validation for dev_setup.sh exit instead of just show
a warning message in start-mycroft.sh
* Added code to allow dev_setup.sh to be run from different
directories successfully
Using a virtual environment is not necessary under Docker, which is a controlled
environment on its own. This uses the semi-standard existence check for ```.dockerenv```
as a way to detect when running in a Docker context.
The start-mycroft.sh and stop-mycroft.sh scripts now automatically
run the enclosure client if the platform is defined as
"mycroft_mark_1".
Also corrected a copy/pasted comment and added a clarifying message during the stop when killing a process.
Add stop support for recent 'python3' change
Necessary after #1703
The -r option eliminates the quotes on output
Clean up the usage of jq, -r removes the quotes on output. Also eliminated the python2 support in the regex ```[p]ython[2|3]``` and dropped the unnecessary ```[p]```.
Verify platform file existence before read
It might not exist on a desktop install.
Verify platform file existence before read
- Remove setup scripts for mycroft-skills-sdk since it's not used anymore
- Rename mycroft-base-setup.in/MANIFEST.in to setup.py and MANIFEST.in
- Remove skill-container, since it hasn't been used or kept up to date since 17.08 and the cli commands to remove and activate skills is easier to work with
==== Environment Notes ====
Small update of the packaging script is needed due to this change
* Fix startup scripts
- change start.sh to the new start-mycroft.sh
- fix startup of messagebus service
- fix startup of enclosure
* Fix running dev_setup.sh as root
Use --allow-root when calling dev_setup.sh
* Correct default directory for user
On the Mark-1 the script would create the mycroft-core repo in /home/root instead of /root when running the script as root
* Skip mimic installation correctly
* Allow doing the github install without root
Running mycroft under root stumbles into issues accessing the audio. allowing unstall under pi user is also more intuitive.
* Change ownership of the webcache
* fix starting/stopping wifi-service
* Fix switching permissions
* Correct startup order
- load messagebus before enclosure
- reboot to workaround audio input issue
* Disable updates when switched to github install
* Fix identity update
identity updates weren't transfered back to the mycroft user since the linked file was overwritten. Now the identity folder is linked instead.
Security researcher w00t pointed out a potential exploit if an user
starts the Mycroft messagebus service on a system with an exposed
network connection with no protection for port 8181. For now the
start-mycroft.sh script will simply display a CAUTION to point out
this concern and remind them to protect themselves with a firewall.
Future versions of the websocket implementation will use encryption
and authentication.
This also adds the possibility to set the environment variable MYCROFT_VENV to separate the virtual env for python3 from the stable version.
Example:
export MYCROFT_VENV=mycroft-18.02
./dev_setup.sh
./start-mycroft.sh all
mycroft-core.sh has a relative-path call (that itself also likes to
be called from inside mycroft-core) that prevents making a clean start
of mycroft from outside the mycroft-core directory. This makes doing
things like adding it as a Ubuntu Unity startup application slightly
messy.
Remove the "restart" option, instead all of the background services
now automatically check to see if they are already running and will
be stopped then restarted if so.
Also fixed a few Codacy complaints.
The default when no parameter is entered is now to display the help.
To start services you can use 'all', or 'debug' to start them followed
but starting the CLI
Reworking the mechanisms used to interact with Mycroft. The old
mycroft.sh utilized the 'screen' command, but now with the CLI it
it easier to use that for viewing log files. Plus there was
confusion between when to use start.sh and when to use mycroft.sh.
Now things are wrapped in the simple and easily discoverable scripts:
* start-mycroft.sh
* stop-mycroft.sh
The start-mycroft.sh allows you to start the background services and
to invoke the various tools (e.g. the CLI, unit tests, etc).
==== Fixed Issues ====
==== Documentation Notes ====
This will require revamp of the getting started guides, etc.
Åke Forslund
Steve Penrod
Julien Kassar
Matthew D. Scholefield
penrods
Rhett Aultman