- the index link ("View this book's table of contents.") which pointed
to http://drupal/book instead of to ... this book's table of contents.
- it doesn't print "previous", "next", "up", "index" if there are no
"previous", "next", "up", "index"
- it doesn't print "index" if "up" /is/ the "index"
- it swaps the order of "up" and "index", ie. "up" is in the row above
"index", between "previous" and "next" now. this is more logical and the
way all tree-browsing apps (python doc, info, ...) do it.
* After having a chat with Dries we have agreed on the logics (or
more specifically he told me how its supposed to work). The queue
module is just a method to make moderation public, but its not
necessarily the only way. This brings the power back.
- Minor interface tweaks (names, case changes, etc).
- Fixed error that would show when you tried to view the profile of a
blocked user.
- Added node_admin_settings() to let users set node defaults.
* Added _nodeapi("conf")
- Modified form_checkbox() to unset options. (Modified the patch from Ax.)
- Removed the _save() hook.
- Fixed a logical flaw in field_get() and field_set() which would break
moderation.
- Moved some admin menus around and cleaned up some typos.
The following modules need updating:
* glossary module
* feed module (Breyten's version)
* mailhandler module
* notify module
* project module
* smileys module
* admin module
* style module
* taxonomy_dhtml module
To avoid unexpected problems menu_add() is deprecated (it will print an
error message when used) and menu() should be used instead.
- Added a _nodeapi hook that in time will replace the _node hook. Trying
to make as few changes to existing code as possible until the new api
is stabilized.
- Modified node_form() to add administration options from other modules
using the _nodeapi hook.
- Modified node_save() to fetch which fields should be saved to the node
table from the _nodeapi hook.
- Moved comment and queue options from node_form() and node_save() to
the modules _nodeapi hooks.
- Removed all instances of '$user->nodes'.
- Committed Moshe's taxonomy patch - minus the node_compact_list() bit. It needs a bit more thought/work. This patch changes the links of taxonomy pages/feeds so update your custom code and themes accordingly!
Themes should now use "taxonomy_link("taxonomy terms", $node)" to get an array of taxonomy term links. The old construct is deprecated and should be changed.
// old theme blob:
if (function_exists("taxonomy_node_get_terms")) {
foreach (taxonomy_node_get_terms($node->nid) as $term) {
$terms[] = l($term->name, NULL, array(), "or=$term->tid");
}
}
// new theme blob:
if (module_exist("taxonomy")) {
$terms = taxonomy_link("taxonomy terms", $node);
}
// old URL:
http://foo.com/index.php?or=1,2
// new URL:
http://foo.com/?q=taxonomy/page/or/1,2
- forum: fixed link to new topic
- forum: new topic shows default forum correctly
- forum: first_new is back; the anchor didn't consider multiple pages
- forum: use standard pager, needed some changes/fixes to pager.inc
- forum: some cleanup
- forum: taxonomy hook
- renamed first_new to simply new
- added an optional parameter to pager_query for the count query
- used the optional count param for paging forum topics
- internal change: moving a topic doesn't duplicate the node anymore but just
changes the forum (term); no change in functionality, and shadow still
works (suggested by Dries). This probably also made some queries somewhat
lighter.
- bug fixed: anonymous users always saw "n (n new)" in the replies column
- updated pager help and moved from _help to phpdoc
- The href target for a new window is "_new", not "new".
- Generating <div> sections within <p> sections is forbidden by the
XHTML standard. Using just the right aligned <div> should be
sufficient and makes XHTML themes possible.
(Prove at http://blog.rompe.org/ )
- While parsing the header of an RSS feed one should be aware that
there may be more <title> tags in subsections and that POSIX regular
expressions are always gready. So make shure we don't get too much.
(If you agree that using PCRE instead of the POSIX ones would be
generally a good idea, then I am willing to make the patch, but for
now I didn't want to mix POSIX and PCRE in one file.)
(Prove at http://blog.rompe.org/index.php?q=import/feed/43 , try
this feed without my patch)
- Some RSS 2.0 feeds don't have a per item <link> section but have the
permalink embedded in the <guid> section. This is not perfectly
correct and the documentation mentions this possibility only in the
examples, but since Dave Winer himself implements it this way it
will happen more than once. So, if there is no link available and
the guid looks like an address, then use that one.
(Prove at http://blog.rompe.org/index.php?q=import/feed/22 , try
this feed without my patch)
- Don't only write eventually new Feed Header information into the
database but also use them immediatly. Reuse the $feed array that is
made for it.
- If a feed doesn't provide per item titles, make shure to not produce
defective markup by cutting the remainder of an entity. Instead of
just cutting off anything behing the leading 30 characters of the
cleaned description, it seems slicker to use up to 40 characters and
split on word boundaries, but not on "&" or ";".
(Prove also at http://blog.rompe.org/index.php?q=import/feed/22 .
This feed will have title tags starting on February 1st, but I
suspect many others without them out there.)
With this patch one could consider Drupals aggregator RSS 2.0 ready.
parent is supposed to exclude terms that are descendants of the current
term, to avoid creating cycles. However, the terms with the first few
term IDs are excluded instead. Patch by JonBob.
or "operation" link.
- Made the main page of the administration section show an overview of all
watchdog entries with such action link.
- Fixed typo in PostgreSQL database scheme.
- rewrote taxonomy_get_tree() for improved performance and cleaner code
- fixed a bug in _taxonomy_term_select() with multiple parents
- added hooks in vocabulary and term insert, update and delete
- fixed a bug in taxonomy_save_vocabulary() (cache_clear_all() was never
called)
- Usability improvement: made the block administration pages use the
new table rendering functions. By the way, the old code rendered
invalid HTML code.
- Usability improvement: saving block changes will now show a status
message.
- Usability improvement: quite a few strings could not be translated.
TODO:
- I think the "enabled" and "custom" field are somewhat confusing.
Suggestions?
the new table rendering functions, made the overview table's HTML
code a lot simpler, tidied up the status messages, and removed some
redundant check_output()s.
the new table rendering functions; access rule table, permission table,
role table, etc.
- Usability improvement: made the permission table translatable. Requested
by Stefan.
- Code improvement: removed some redundant check_output()s.
- Usability improvement: made the user administration pages use the
new table rendering functions.
- Usability improvement: added pager support to the user administration
pages.
- Code improvement: fixed code indentation.
- Usability improvement: made the comment administration pages use the
new table rendering functions.
- Usability improvement: new comments will also be marked as new in the
administration pages.
- Code improvement: changed the comment administration pages to use the
default pager instead of a custom pager.
- Bugfix: "yes" and "no" were not translatable.
- $ diffstat comment.patch
comment.module | 45 +++++++--------------------------------------
1 files changed, 7 insertions, 38 deletions
TODO:
- Comment module: remove other custom pagers in the comment module.
- Pager/database: it would be nice if the functions db_query_range()
and pager_query() would have the same semantic as db_query(); that
is, using a variable number of arguments that are automatically
escaped/checked.
- Removed a broken filter which caused some line breaks to be stripped.
- It is now possible to create short teasers and the '---' delimiter can
be outside the allowed range but not more than a factor two.
- Fine-tuned teaser extraction in absence of the '---' delimiter.
when the taxonomy module is disabled. Reported by kyvinh.
- Improved forum documentation. Modified version of Kjartan's patch.
TODO:
- Updating a forum topic through the admin interface, results in a
duplicate (and incomplete) forum topic. Also note that "Leave
shadow?" function is annoying; the status of the checkbox is lost
when clicking preview. Also, I think it should be disabled be
default.
- consistency in return order
(http://list.drupal.org/drupal-devel/2002-December/009522.html)
- new functions: taxonomy_get_term_by_name() and
taxonomy_get_vocabulary_by_name()
- caches are flushed after vocabulary and term edit/delete; this should
avoid problems like forums not up to date
- page_header() now adds Last-Modified and ETag http headers.
- When running PHP as an Apache module page_header() will check the HTTP
headers for conditional gets, and will only push the content when it
fails. (Works for html and xml pages as they are all cached). Note:
this is a PHP limitation, so until PHP makes it work for other web
servers this won't work for them.
- Added created field to cache database to hold the timestamp when the
cache was created.
- Changed cache_get() to return an object with ->data and ->created.
- Update forum and locale modules.
+ Removed the "post content" permission and replaced it by more fine-grained permissions such as "maintain static pages", "maintain personal blog", "maintain stories", etc.
o Usability improvements to teasers:
+ Teaser forms are no more. Teasers are extracted automatically but can also be instructed using a delimiter "---". Furthermore, when a post it too short for a teaser, the user won't be bother with teaser stuff anymore.
+ Added an option to set the teaser length, or to disable teasers all together.
+ When previewing a post, both the short (if any) and the full version of a post are shown. This addresses a common complaint; for example, when writing a book page there was no way you could preview the short version of your post.
+ Forum posts can be teasered now. This is particularly helpful in the context of drupal.org where we promote forum topics.
o Bugfix: replaced all PHP short tags (<?) with long tags (<?php).
o Bugfix: removed hard-coded dependence on comment module.
o Bugfix: when the queue module was disabled, it was not possible to approve updated book pages.
o Bugfix: applied modified version of Marco's node_teaser() fix.
- Fixed a bug with get_tree and multiple parents.
- Fixed 3 wrong caches (this will avoid some queries).
- Extension to taxonomy_term_count_nodes() (feature #159): "currently
calling taxonomy_term_count_nodes() returns the amount of nodes in each
term, would it be possible to extend this function so that it would only
return the amount of nodes of a certain type under each term.
- Confirm deleting a vocabulary or a term (requested by Moshe).
- Use form_weight() in vocabulary and term forms.
- After submitting a term, we end up at the term page (requested by Moshe).
- Added status messages when adding, editing and deleting vocabularies and
terms
- Minor clean ups
* Made sure all modules use the same link delimiter; "|" for now.
- Changed cache API.
- Fixed caching bug in comment.module. Odd this hasn't been reported yet.
- Fixed caching bug in forum.module.
- Fixed caching bug in system.module.
- Fixed caching bug in block.module.
- Simplified caching support in forum.module thanks to improved cache API.
- Adds more title info to each line in the block: now showing author and number of comments in a unified way accross all modules.
- Added pager support to blog module.
- Blog module now uses standard node view, not its own hack which is more theme friendly (http://www.drupal.org/node.php?id=133).
- Blog module now uses node_feed() function for making RSS feeds. since all feeds are now consolidated, a couple tweaks to node_feed() and format_rss_* would lead us to support new RSS formats easily.
- Fixed bugs where unpublished entries were displayed.
now be changed on the book overview page in the admin page. Editing the
"Drupal handbook" (100+ pages?) on drupal.org became a time-consuming
process, hence these improvements ...
"This patch to theme.inc adds the ability for modules to insert HTML
into the <HEAD> section of all web pages. The additional modules in
this directory [ed: blog.module] demonstate possible uses for the
new _head() hook."
"This is a simple, generic pager for Drupal-CVS. It is designed to be
easily themeable and expandable. The code is highly-commented to
enhance readability."
"Pagers are constructed by combining the provided pieces (all of which
can be easily modified to display the text or image you prefer) into
your custom pager."
* Statistics module fixes by Jeremy:
- removed superfluous check for existence of watchdog()
- saving changes in admin page displays status and returns same page
- no longer return 1971/01/01 in "view statistics" table
- switched from "!=" to "<>" in SQL queries for ANSI-SQL compliance
- switched from "MAX(timestamp) as timestamp" to "MAX(timestamp) as
max_timestamp" moving towards ANSI-SQL compliance.
* Added a "theme_item_list" function to format itemized lists. Also
changed a couple of modules to take advantage of it. Makes for a
more consistent UI.
- Modules and themes now use the same functions to find and administer
files.
- Modules can now be placed in sub-directories.
- Theme descriptions can no longer be edited. This will be handled by
Dries' theme_conf patch.
- Update required to keep old modules enabled.
- addressed one of Ax's issues:
> = there are quite some "$output .= " which should be "$output
> = "
> (whenever $output hasn't been initialized before)
- combined a few lines that could be accomplished in one step
- changed "!" to "." in comment "This site is currently
sustaining..." as if a low number, it looks silly.
correctly. Patch by Stefan.
- Added missing t() functions and improved the %x directives for
better readability and to ease the process of translating your
site. Patches by Stefan.
- Made two small additions to the 'code-clean.sh' script; it will
now remove patch related junk.
NOTES:
- I removed the · related bits. Let's tackle these later on
in a separate patch after we got some sort of consensus.
- I removed the 'module_exist("drupal")' check in the user module;
I *think* it is incomplete and therefore incorrect.
- Stefan, try using quotes in your translations and check whether
everything still works. Example: translate the "Create account"
button to "Create \"haha\" account \'hihi\'" and see if you can
still create new accounts. Maybe automate this using a quick
hack in the locale module ...
+ Blocks are not longer called if not rendered: major performance
improvement.
+ Fixed some bugs (preview option was broken, path option was broken).
+ Removed "ascii"-type blocks.
+ Added permission to for "PHP blocks"
+ ...
NOTES:
+ You'll want to run "update.php":
ALTER TABLE blocks DROP remove;
ALTER TABLE blocks DROP name;
+ You'll want to update your custom modules as well as the modules in
the contrib repository. Block function should now read:
function *_block($op = "list", $delta = 0) {
if ($op == "list") {
return array of block infos
}
else {
return subject and content of $delta block
}
}
changes include:
* a couple of coding style changes, renamed some "stats" into
"statistics", etc.
* removed the "Who's online" block from the user module.
* added db_affected_rows() to the resp. database abstraction
layers and made the statistics module use db_affected_rows()
instead.
* added update logic to "update.php".
Dries Buytaert
Kjartan Mannes
Steven Wittens