aboutsummaryrefslogtreecommitdiff
path: root/libindicator
Commit message (Collapse)AuthorAgeFilesLines
* indicator-ng: simplify flow in initable_initLars Uebernickel2013-01-301-21/+20
|
* indicator-ng: use strlen instead of hard coding the lengthLars Uebernickel2013-01-251-1/+1
|
* indicator-ng: use indicator_image_helperLars Uebernickel2013-01-253-8/+27
| | | | | | | | gtk_icon_set_from_gicon doesn't scale rectangular icons correctly. This adds indicator_image_helper_update_from_gicon() and makes refresh_image() set a broken image instead of erroring out when an icon couldn't be found.
* indicator-ng: lazily allocate entry.label and entry.imageLars Uebernickel2013-01-251-7/+27
| | | | | Most indicators only need one of those.
* indicator-ng: require header item to have x-canonical-type setLars Uebernickel2013-01-251-15/+38
|
* indicator-ng: check return value of g_menu_model_get_item_attributeLars Uebernickel2013-01-241-3/+5
| | | | | If it returns false, we'd use uninitialized memory.
* indicator-ng: document error conditions in menu_changedLars Uebernickel2013-01-241-0/+4
|
* indicator-ng: save unnecessary allocations by using "&" in g_variant_getLars Uebernickel2013-01-241-8/+4
|
* indicator-ng: show broken image when g_icon_for_string returns NULLLars Uebernickel2013-01-241-17/+18
|
* indicator-ng: use an IndicatorObjectEntry internallyLars Uebernickel2013-01-241-83/+26
| | | | | We needed the entry anyway (that's what indicator_ng_get_entry was for).
* indicator-ng: set name hint to the value of the service file's "Name" fieldLars Uebernickel2013-01-231-8/+23
|
* indicator-ng: use base->get_entries to get the invisible ones, tooLars Uebernickel2013-01-231-1/+1
|
* indicator-ng: auto start service if it's not runningLars Uebernickel2013-01-231-1/+1
|
* indicator-ng: properly unset action group when the service disappearsLars Uebernickel2013-01-221-1/+1
|
* Make sure indicator-ng.h is installedLars Uebernickel2013-01-211-3/+5
|
* indicator-ng: add gettersLars Uebernickel2013-01-212-0/+20
|
* indicator-ng: add indicator_ng_new_for_profileLars Uebernickel2013-01-212-3/+18
|
* indicator-ng: fix crash (tried to free a string with g_object_unref)Lars Uebernickel2013-01-211-1/+0
|
* indicator-ng: always set an accessible description to avoid imminent warningLars Uebernickel2013-01-211-0/+6
|
* indicator-ng.h: use local includeLars Uebernickel2013-01-211-1/+1
|
* Only build and use IndicatorNg in the gtk3 buildLars Uebernickel2013-01-181-1/+5
|
* Add IndicatorNgLars Uebernickel2013-01-183-1/+468
| | | | | | | IndicatorNg is an indicator object that reads an indicator service file and watches the bus for a corresponding service to appear. It turns the menus and actions exported by the service into an indicator entry.
* No newline at end of fileCharles Kerr2012-09-111-1/+2
|
* remove the cloak/decloak widget logicCharles Kerr2012-09-111-64/+3
|
* don't include build-time filenames in comments of the enum header filesEvgeni Golov2012-06-102-3/+3
| | | | | | | If we do, the header files may be different on each build. This is especially harmful when building multiple times for multiple architectures and expecting the files to be identical.
* indiator_service_manager_new*()'s name arguments were gchar*, make them ↵Charles Kerr2012-04-222-4/+4
| | | | const gchar*
* in indicator_image_helper_update(), test the image argument with ↵Charles Kerr2012-04-111-1/+1
| | | | GTK_IS_IMAGE before using it.
* Fix spelling/grammar errors in an error message.Charles Kerr2012-04-021-1/+1
|
* Setting again the variables as constMarco Trevisan (Treviño)2012-04-011-2/+2
|
* Indicator Service Manager: we don't have to free values if got with &s from ↵Marco Trevisan (Treviño)2012-04-011-3/+0
| | | | | | variant See http://is.gd/EyhAhP
* merge lp:~htorque/libindicator/fix-small-leak to fix a GError memory leak ↵Charles Kerr2012-03-261-0/+1
|\ | | | | | | reported in lp bug #965459
| * Free GError.Hernando Torque2012-03-261-0/+3
| |
* | fix minor memory leak when the proxy name changesCharles Kerr2012-03-161-2/+5
|/
* Copying the name before using it to remove as it seems that the unwatch can ↵Ted Gould2012-03-021-1/+3
| | | | cause the name to be free'd in some cases. Odd, but valgrind can find it
* Support new FD.o Desktop ActionsTed Gould2012-03-021-15/+95
|\
| * Get the right group in exec too! Ted Gould2012-02-271-1/+15
| |
| * Handling the different group names when getting the name as wellTed Gould2012-02-271-1/+15
| |
| * Creating a warning for the legacy supportTed Gould2012-02-271-0/+1
| |
| * Adding a value for whether it should use the TargetEnvironment, I'm removing ↵Ted Gould2012-02-271-5/+4
| | | | | | | | the warning because we don't care to fix old stuff anymore
| * Use the appropriate list of nicks and groups to find our data nowTed Gould2012-02-271-2/+24
| |
| * Figure out what kind of actions we haveTed Gould2012-02-271-1/+17
| |
| * Adding a state variable to track what kind of actions we haveTed Gould2012-02-271-0/+11
| |
| * Changing our defines around to make sure to mark things the way that we ↵Ted Gould2012-02-271-11/+14
| | | | | | | | think about them.
* | silence Coverity warning "PW.BRANCH_PAST_INITIALIZATION - CID 10615" ↵Charles Kerr2012-02-291-2/+1
| | | | | | | | | | | | | | reported in Bug #939061 Coverity is unhappy that the local variable "priv" is uninitialized if one of the goto branches is followed in this function. Since we only use priv once, a fairly clean way of sidestepping this warning is removing the local variable "priv" and instead using INDICATOR_OBJECT_GET_PRIVATE()'s return pointer directly.
* | tweak: use g_clear_object() instead of if-not-null-g_object_unref()Charles Kerr2012-02-291-6/+2
|/
* Removing unenforcable constsTed Gould2012-02-222-4/+4
|
* indicator_object_dispose: remove unneeded list of entriesLars Uebernickel2012-02-131-5/+1
|
* indicator_object_dispose: use _set_visible instead of reimplementing itLars Uebernickel2012-02-131-12/+2
|
* indicator_object_set_visible: no need to check for ENTRY_INITLars Uebernickel2012-02-131-4/+2
|
* indicator_object_set_visible: only emit entry-{added,removed} when ↵Lars Uebernickel2012-02-121-1/+8
| | | | | | | | | | | visibility actually changed Emitting those signals is semantically incorrect. It also confused indicator-loader and the panel, because they are not checking whether an entry is already present in their entry-added handlers. This led to gtk warnings ("widget already has a parent") and strange effects, such as the same menu being added multiple times into the panel.