GitLab

The resume composites were broken when editing a resume (although they were working in the views). Fixed.

Make sure we check for the correct langconfig.php, regardless of whether we're retrieving a plugin string or not.

Fixes a gazillion warnings to the logs when using language packs other than English.

Eduforge Bug 1716

Namely: the string saying how many results there are, and the string that says there are no artefacts to choose from.

Also removed styles for the old administer templates screen that I found along the way.

Those exceptions are safe to translate, other exceptions can't easily be translated without testing, as you have to make sure everywhere they should be used the translation infrastructure is working.

As it is, if the translation code throws an exception, everything will probably go to hell.

I was actually looking to fix a bug about it displaying the currently logged in user in search results, but there's a change on the forums branch that will allow this, and I don't want conflicts.

Also noted in undelete_user what has to be done to undo this change, but I haven't implemented it because there's nowhere in Mahara where undeleting a user is possible in the UI anyhow.

Ensure that when a user is deleted, suspended or has their account expired, they really are unable to log in.

Not sure for how long this has been broken, but it seems that deleted users could browse around the site just fine. They don't show up in search results but you can still view their profile page...

Now the items are layed out in a grid to the right of the searcher, which saves some vertical space and has more fluidity to the layout.

Also, the calendar elements have from/to text, which is a small violation of the language text freeze but makes the purpose of the calendar items for each element much clearer.

I'd like to add more text to this page as an explanation, though that would be a larger violation.

Changed a smarty() call to smarty_core(), using the full object for a form element template is overkill.

The fix basically works by adding a hidden submit button at the very beginning of the form. Its "name" is the same as for whatever current action is being performed. Thus, when enter is pressed, the same value is passed through.

In javascript mode, this form element's name attribute is poked every time a new element is configured to be correct for that. There is also some special event handling for the search tab of the artefact chooser, so it does not trigger a form submit - just a 'click' event on the correct button.

Make sure that errors don't occur if pagination DOM nodes aren't available (similar to previous commit, but for the paginator itself)

Handle the case when there is no pagination related DOM stuff available - as happens when there are no results.

Fixes: If the artefact chooser has no tabs, then every single link on the page will be wired up to attempt to retrieve more data via AJAX for the artefact chooser.

Fixed by checking that there are actually tabs first.

Together with the previous commit, fixes #1720

Together with the next commit, solves #1720

Prevent a fatal error caused by adding the entireresume blocktype then visiting your resume for the first time.

The 'contactinformation' blocktype is attempted to be added to the entireresume blockinstance that has been created, causing badness as the blocktype hasn't actually been committed yet. Move the part of the commit where it's added to the database before the part where it's added to any views.

Also make sure that contact information isn't shown in a view as just a <table></table> if it has no content.

Updated search code for solr 1.2. Now send correct Content-type header, and parse responses the 1.2 way.

Also, store tags as multivalued arrays.

Also put in a check for no results being returned.

Make the tags field 'textTight' to make it a more restrictive match, closer to how the internal tag search now works.

Also made it multivalued.

Fixes bug #1332.

- Only create the directory in the postinst script
- Delete the directory in the postrm script

I got confirmation from the author directly:

From: James Stewart <james@jystewart.net>
Date: Sat, 24 Nov 2007 12:03:20 +0000
To: Francois Marier <francois@debian.org>
Subject: Re: Distributing XML_Feed_Parser

Hi Francois,

Thanks for getting in touch. I had indeed meant to license it under the
LGPL but must have made a mistake when setting up the package.xml file.
I've just committed a new version to CVS with the LGPL listed as the
license, but guess the package page won't update till I release a new
version of the package.

I'm slowly working on a new version of the package but it'll be some time
before I get it ready for release.

James.

Move mahara cron log files to /var/log/mahara/ so they can be kept separate from other content in /var/log

Turns out I got the table_column call wrong also. Needs to, like with everything else in xmldb, fully respecify the column. Now the view migration works in MySQL