summaryrefslogtreecommitdiffstats
path: root/NEWS
Commit message (Collapse)AuthorAgeFilesLines
* 3.153.15Thierry Vignaud2015-05-181-0/+4
|
* 3.143.14Rémi Verschelde2015-02-221-0/+5
|
* 3.133.13Rémi Verschelde2015-02-161-0/+3
|
* Remove apparently unused get_product_info subroutineRémi Verschelde2015-02-161-0/+3
|
* Update headers and FSF contact addressRémi Verschelde2015-02-161-0/+2
|
* 3.123.12Anne Nicolas2015-02-091-0/+4
|
* 3.113.11Thierry Vignaud2015-01-021-0/+4
|
* 3.103.10Thierry Vignaud2014-02-041-0/+2
|
* fix crashing if there's not notification daemon (mga#12416)Thierry Vignaud2014-02-041-0/+1
|
* fix gtk2 to gtk3 migration: XWINDOW -> get_xid (mga#12527)Thierry Vignaud2014-02-031-0/+1
|
* restore translations from commit 06c0d4a37aec130691c55e13937bc688bf5e6cb1Thierry Vignaud2014-02-031-0/+1
| | | | | they were lost due to stupid/gratuitous breakage in commit d67b95af569556af4a4b333e1707fdbc34c447e5
* include strings from mgaapplet-update-checker & mgaapplet-upgrade-helperThierry Vignaud2014-02-031-0/+2
|
* 3.93.9Colin Guthrie2014-01-251-0/+2
|
* Do not exec pkexec but run it so it has a parent process to fix double fork ↵Colin Guthrie2014-01-221-0/+3
| | | | issues mga#11184 mga#12364
* 3.83.8Thierry Vignaud2014-01-221-0/+2
|
* block CHLD signals on startupThierry Vignaud2014-01-221-0/+3
| | | | | | in order to prevent glib-threading segfaults (mga#10289), just block the CHLD signal during the window where glib create threads behind our back (RT-120951)
* 3.7.13.7.1Thierry Vignaud2014-01-221-0/+2
|
* better fix for mga#12280 while fixing exit reported by ColinThierry Vignaud2014-01-221-0/+2
| | | | | | | | | | | | | | | | | | previous fixes for mga#12280 resulted in Colin reporting mgaapplet exiting (not segfaulting): "When updates a bubble pops up: - If you click on the background of this bubble, mgaaplet exits - If you click on the "Install Updates" button, it pops up the password dialog. Without touching anything, about 1s later the dialog seems to be replaced with another that says "Sorry, that didn't work, please try again". When this dialog appears, mgaapplet exits." So let's clean the pile of fixes and: - have only _one_ gtk+ main loop - creating _one_ hidden notification before entering the main loop - never destroy it, but use >close + >clear_actions() then ->update() prior calling >show() when one need to show a notification again
* Update gnome-shell code to exit message tray mode also.Colin Guthrie2014-01-221-0/+3
| | | | | | | | The old method worked on older gnome-shells, this new method works on newer gnome-shells (see bgo#706783) Ideally neither hacks should be necessary and maybe (eventually) we can remove them, but for now...
* 3.63.6Thierry Vignaud2014-01-171-0/+2
|
* workaround X11 hanging (mga#12280)Thierry Vignaud2014-01-171-0/+3
| | | | | | | | | | | | | | | | | | | | | | | Since commit 95b9cd06f14a9817090584d72830df870c591acc, we run a gtk main loop after displaying a notification, else actions when clickong notification buttons are ignored by gtk+/libnotify However, if the notification is not manually closed, we never exit this main loop. In that case, gtk+ fails with: (mgaapplet:9060): Gtk-CRITICAL **: gtk_window_set_accept_focus: assertion 'GTK_IS_WINDOW (window)' failed from: data=<optimized out>, destroy=0x0, button=3, activate_time=5407876) at gtkmenu.c:1613 And X11 is stuck. As a workaround, since libnotify offers no way to be notified when notification is automatically closed, just add a timeout for exiting the main loop. At worse, X11 will be stuch only 5 seconds.
* 3.53.5Thierry Vignaud2014-01-081-0/+2
|
* delay setting SIG_CHLD handler (mga#10289)Thierry Vignaud2014-01-081-0/+3
| | | | thus fixing segfault on startup
* 3.43.4Thierry Vignaud2013-12-301-0/+2
|
* drop support for mdv 200[89].xThierry Vignaud2013-12-301-0/+1
| | | | | basically reverting commit 56bcaff85b7b8a98ffdf67118a4df873fe301882 from Nov 24 2009 ("(configure) make it work on 2008.x & 2009.x")
* fix two Entry pointing to the same reference (mga#12148)Thierry Vignaud2013-12-301-0/+1
| | | | | | | | bug introduced in commit 56bcaff85b7b8a98ffdf67118a4df873fe301882 on Nov 24 2009 (was: "(configure) make it work on 2008.x & 2009.x") it wasn't visible in newer distro as we were relying on the other code path
* fix missing arguments in widget creation test (mga#12148)Thierry Vignaud2013-12-301-0/+3
| | | | | | regression introduced in commit 6c1fa5cd83b0d46528ee8cecda42f9b1adca0b32 on Nov 18 2013 ("use (my|u)gtk3 instead of *tk2") when switching from gtk2 to gtk3
* 3.33.3Thierry Vignaud2013-12-211-0/+2
|
* fix doing nothing when clicking on notifications buttons (mga#12041)Thierry Vignaud2013-12-211-0/+1
|
* split mgaapplet-update-checkerThierry Vignaud2013-12-211-0/+3
| | | | | | since we cannot prevent glib/gtk to spawn threads behind our back, we can at least try to prevent segfaults due to mixing threads with secular forks by exec()ing immediately
* 3.23.2Thierry Vignaud2013-12-161-0/+2
|
* fix crashing when displaying about dialog (mga#12009)Thierry Vignaud2013-12-161-0/+2
| | | | set_authors() needs an array ref with Gtk3
* 3.13.1Thierry Vignaud2013-12-061-0/+2
|
* Gtk3::Notify->new => Gtk3::Notify::Notification->new (mga#11897)Thierry Vignaud2013-12-061-0/+2
|
* 3.03.0Thierry Vignaud2013-12-051-0/+4
|
* 2.842.84Colin Guthrie2013-11-171-0/+2
|
* polkit: Only use pkexec when not already rootColin Guthrie2013-11-171-0/+2
|
* 2.832.83Colin Guthrie2013-10-131-0/+2
|
* Add urpmi.update polkit wrapper.Colin Guthrie2013-10-131-0/+2
| | | | | | | | | We like to be able to run this command without any user interaction when checking for package updates. It's somewhat awkward to keep the policy and the binary in separate repositories, but keeping this split will prevent pulling in mgaonline during priority updates section of the distro upgrade.
* 2.822.82Colin Guthrie2013-10-121-0/+2
|
* Use polkit for authorisation (mga#11125)Colin Guthrie2013-10-121-0/+1
|
* Do not handle priority upgrades specially. This is now fixed in urpmi itselfColin Guthrie2013-05-191-0/+1
|
* Do not pass invalid --nocheck option to urpmi.updateColin Guthrie2013-05-191-0/+2
|
* 2.812.81Colin Guthrie2013-05-151-0/+2
|
* Add missing news about previous commitsColin Guthrie2013-05-151-0/+2
|
* Enable the 'download all' links for everyone (mga#8157)Colin Guthrie2013-05-141-0/+1
|
* Do not use urpm code to check the folder as this is run as unprivileged user ↵Colin Guthrie2013-05-131-0/+1
| | | | (mga#6083)
* Ensure nonfree/tainted media is kept enabled on distro upgrade.Colin Guthrie2013-05-131-0/+2
| | | | | | | | Patch from Maarten Vanraes and Dave Hodgins and tweaked by me for trunk. mga#6061
* 2.802.80Colin Guthrie2013-05-101-0/+2
|
* Add support for upgrades that need some form of preparation before continuing.Colin Guthrie2013-04-281-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | The basic principle of this is to do the following: 1. Check the releases webservice API to see if the distro version 'needs_preparation'. 2. If it does, check for a file /var/lib/mageia-prepare-upgrade/state. If it says 'ready' then we are ready and the user is not questioned further. 3. If it does not exist or says anything else, the user is prompted to install the package 'mageia-prepare-upgrade'. 4. After package installation, check the state file again as this may be all that is required. If the state file does not yet say 'ready', then display the text from the /usr/share/doc/mageia-prepare-upgrade/README.prepare file (todo: add i18n) 5. Assuming that further action is required by the user (e.g. rebooting to convert the filesystem), then the upgrade helper will exit. 6. After the user has taken the action required to do the preparation, the state file should contain 'ready'. The next time the distro upgrade dialog shows, the preparation checks will pass and the user can continue. TODO: Handle i18n in the README.prepare file. v1: First Revision v2: Check immediately if an upgrade state file exists (don't wait 5 mins) Ensure config file is parsed (ensures the testing variable is read) Show user-visible messages when bailing out due to unexpected errors