Robert Sesek [Sun, 8 Apr 2007 01:22:20 +0000 (01:22 +0000)]
r1538: Rebuild all the caches in install/upgrade13.php just to be safe because I think there may be some bad data from pervious installs... and it can't hurt (actually, we may want to do it after every upgrade)
Robert Sesek [Sat, 7 Apr 2007 02:00:13 +0000 (02:00 +0000)]
r1533: Apparently some systems don't like our pass-by-reference use of ProcessBugDataForDisplay(), so change it to be used with an assignment operation instead. Closes bug://report/78.
Robert Sesek [Mon, 2 Apr 2007 19:06:29 +0000 (19:06 +0000)]
r1528: In install/installer.php:
- Move the data caching system up to be right after we load settings and default data
- Call build_usergroups() right before we insert the new admin user, removing the problem of the usergroup datastore not being built for build_assingedto()
Robert Sesek [Mon, 2 Apr 2007 18:32:10 +0000 (18:32 +0000)]
r1526: In includes/class_notification.php#fetch_user_cache(), we no longer want to clear the roles[] array because it would actually prevent emails from being sent due to the way we use set_bug_data() in editreport.php
Robert Sesek [Wed, 28 Mar 2007 00:22:21 +0000 (00:22 +0000)]
r1524: In BugAPI::set(), we can't use our pretty call_user_func_array() logic, but rather we now have to do a physical call to parent::set() because of stupid PHP4
Robert Sesek [Tue, 27 Mar 2007 02:28:12 +0000 (02:28 +0000)]
r1523: In passwordreset.xml.tpl, we need to enclose the trackertitle variable in curly braces so it doesn't come out as "Array['trackertitle']" in actual emails
Robert Sesek [Mon, 19 Mar 2007 17:39:47 +0000 (17:39 +0000)]
r1512: In the approve users section of admin/usergroup.php, we were still using the mailing system from ISSO 2.0. Switched to using template-based emailing (which requires us to load the template system in that one file). Added templates/email/accountapproved.xml.tpl
Robert Sesek [Mon, 19 Mar 2007 00:04:28 +0000 (00:04 +0000)]
r1510: In draw_cancel() [templates/global.js] we can't use double quotes for window.location because that wouldn't mesh well with being placed inside an onclick attribute... in fact I dare say it would down right break it.
Robert Sesek [Sun, 18 Mar 2007 18:01:33 +0000 (18:01 +0000)]
r1501: We can actually simplify Logging::compare_arrays() by removing the array_diff() business because it really isn't necessary it turns out and was actually responsible for adding a whole bunch of useless logs to the system
Robert Sesek [Sun, 18 Mar 2007 17:47:05 +0000 (17:47 +0000)]
r1500: In showhistory.php:
- Enable multi-group support for custom fields
- Do not show custom fields in the history if users don't have permission to see them
- When we have a useless log (the very first empty checks at the top of the loop), kill them with a database query
Robert Sesek [Sun, 18 Mar 2007 17:35:45 +0000 (17:35 +0000)]
r1499: In editreport.php:
- Only set the standard bug fields if we have permission to change them (thus data isn't cleared when we don't have permission to)
- When setting the bug data for the NotificationCenter, use array_merge() to combine objdata[] and values[] so it doesn't look like we cleared all the fields if we didn't set them
Robert Sesek [Sun, 18 Mar 2007 04:50:06 +0000 (04:50 +0000)]
r1493: In admin/product.php, move the page_start() call in the addversion branch to be after the check for invalid products so we don't end up with a small display bug if the error is thrown
Robert Sesek [Sun, 18 Mar 2007 04:45:39 +0000 (04:45 +0000)]
r1492: Apparently MySQL doesn't think that NULL values should be included in !parentid which causes not all the products to be displayed in the product-level permissions
Robert Sesek [Sun, 18 Mar 2007 04:37:55 +0000 (04:37 +0000)]
r1490: Fixing a SQL error in admin/usergroup.php that would occur when inserting a new usergroup and trying to INSERT INTO the permissions table for custom fields
Robert Sesek [Sun, 18 Mar 2007 01:26:46 +0000 (01:26 +0000)]
r1486: Adding API support for multiple usergroups in UserAPI. Because we verify in the API, we no longer need to call array_strip_empty() in FetchUserPermissions()
Robert Sesek [Sun, 18 Mar 2007 01:22:03 +0000 (01:22 +0000)]
r1485: - Adding base support for multiple user groups: schema changes, created FetchUserPermissions(), and an improved can_perform()
- In ProcessBugDataForDisplay() we can use less calls to can_perform when checking canviewhidden* permissions
Robert Sesek [Sat, 17 Mar 2007 19:36:03 +0000 (19:36 +0000)]
r1479: Remove the need for $stylevar[alt_color] by using the altcolor CSS class by changing the values swapped when calculating rows and using trackerhome_bits.tpl
Robert Sesek [Sat, 17 Mar 2007 19:16:04 +0000 (19:16 +0000)]
r1474: We no loner need the stylevars for cellpadding and cellspacing in tables because we can use CSS properties to achieve the same effect. Removing all the long <table> declarations and just replacing them with styling table with the .table class
Robert Sesek [Sat, 17 Mar 2007 18:42:26 +0000 (18:42 +0000)]
r1473: Because we didn't understand the return val of eval() all our notifications were blank. Fixing by moving all the eval() statements to not be in the assignment operation in NotificationCenter.
Robert Sesek [Sat, 17 Mar 2007 18:25:02 +0000 (18:25 +0000)]
r1471: Fix for bug://report/69: In install/global.php, move the loading of the includes/language.php to be higher up so that the gettext emulation system is enabled sooner
Robert Sesek [Sat, 17 Mar 2007 16:08:45 +0000 (16:08 +0000)]
r1470: Fix two bugs with searching that would:
- Delete the current search if it were ever to be rerun
- Not update the timestamp on cached searches which would cause them to be rerun every time
Robert Sesek [Sat, 17 Mar 2007 15:52:02 +0000 (15:52 +0000)]
r1468: In mailtest.php:
- Check the validity of email addresses before sending
- Make the sending from a non-existent address more effective and meaningful
Robert Sesek [Fri, 16 Mar 2007 02:27:22 +0000 (02:27 +0000)]
r1466: Fixing a bug in NotificationCenter::send_new_comment_notice() where we weren't getting the userid because we were accessing strings and not the $userid variable
Robert Sesek [Sun, 11 Mar 2007 23:20:53 +0000 (23:20 +0000)]
r1449: One last change for class_sort.php that now will make _processDataForColumn() work by changing 'userid' to 'reporter' in step with the previous change from r1444
Robert Sesek [Sun, 11 Mar 2007 22:56:20 +0000 (22:56 +0000)]
r1444: In all other instances for sorting, the sortby key for reporter is reporter, not userid, so have includes/premissions (and thereby $bugsys->columns and $bugsys->columnNames) reflect that
Robert Sesek [Sun, 11 Mar 2007 21:44:23 +0000 (21:44 +0000)]
r1440: - Fixed a huge bug in Authentication::_syncBugdarUser() that would use the authid as the user ID in the Bugdar database as the condition when syncing
- Only call Authenticatoin::clearCookies() in Autthentication::authenticateCookies() if the cookie data fails to verify, and not in other cases