| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The entire rationale for adding this _noatime function was fairly
bogus.
See https://bugzilla.gnome.org/show_bug.cgi?id=680326
But the Linux kernel doesn't permit non-root users to use it for
root-owned files - which is the default for real world GNOME
installations. We end up doing twice the syscall traffic with no
benefit at all.
The patch does work in jhbuild, but honestly the relatime handling is
generally OK for this, and where it's not, change your system to mount
O_NOATIME.
https://bugzilla.gnome.org/show_bug.cgi?id=708453
|
|
|
|
| |
g_strdup already handles NULLs just fine.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When GnomeRROutputInfo is normally created, we strdup the
vendor/product/serial strings that we read from EDID. When it's
finalized, we free them as well.
When we copy the output info, we actually copy the raw struct, and then
selectively strdup members that we care about. When the copy is freed,
we then free the pointer that we dup'd from. When the original output
info is freed, this leads to a double-free. Due to coincidences, on x86,
this doesn't lead to a crash, but on ARM, the heap is completely
corrupted.
To prevent this, when we copy the output infos, strdup them from the
original output as well. We really should clean this API up so that it's
not so awfully RandR-y in style, since it's obvious that the APIs we
have aren't great and really aren't what we need.
https://bugzilla.gnome.org/show_bug.cgi?id=742569
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=742111
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=742111
|
|
|
|
|
|
|
|
|
|
| |
Fix one unusual code path where the GnomeDesktopThumbnailFactory’s mutex
is not unlocked before the function returns. As well as the usual
deadlock issues, this would lead to an abort on shutdown with recent
GLib changes, as g_mutex_clear() would be called on the locked mutex,
which GLib now (correctly) complains about.
https://bugzilla.gnome.org/show_bug.cgi?id=740666
|
|
|
|
|
|
|
|
|
|
|
| |
thumbnail_path did not include the actual filename in its result after
calculating it. This breaks things in save_thumbnail which assumes
receiving a file path rather than a directory path.
See original code before clean ups at:
https://git.gnome.org/browse/gnome-desktop/tree/libgnome-desktop/gnome-desktop-thumbnail.c?id=f2d25fa7d6cdaba03679f0fc0921fc5f81bd3944#n1060
https://bugzilla.gnome.org/show_bug.cgi?id=684026
|
|
|
|
|
|
|
|
| |
g_mkdir_with_parents returns 0 on success or if the directory already
exists, which caused the code to go into the failure path rather than
generating the thumbnail.
https://bugzilla.gnome.org/show_bug.cgi?id=684026
|
|
|
|
|
|
| |
Another cleanup in preparation for a new "simple" thumbnail API.
https://bugzilla.gnome.org/show_bug.cgi?id=684026
|
|
|
|
|
|
| |
Put the path calculation code in one spot.
https://bugzilla.gnome.org/show_bug.cgi?id=684026
|
|
|
|
|
|
|
|
| |
Add DSI to the output name in
_gnome_rr_output_name_is_builtin_display() to enable screen
rotation on Dell Venue 8 Pro.
https://bugzilla.gnome.org/show_bug.cgi?id=740289
|
| |
|
|
|
|
| |
This reverts commit 786336cc7cfaa8e4a0acafe5b9868950e538b740.
|
|
|
|
|
|
| |
They aren't used, and don't work on wayland systems anyway.
https://bugzilla.gnome.org/show_bug.cgi?id=739463
|
|
|
|
| |
it's not used.
|
|
|
|
|
|
|
|
|
|
| |
At least OpenBSD's libc doesn't accept 'utf8' as a locale codeset
suffix but does accept 'UTF-8'. Since GNU libc accepts both suffixes
let's use the one which works on a broader set of systems.
Thanks to Stefan Sperling and Antoine Jacoutot.
https://bugzilla.gnome.org/show_bug.cgi?id=710412
|
| |
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=735974
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=736482
|
|
|
|
|
|
|
|
| |
Layout->xkb_name isn't unique so we can't use it as a key in the hash
table. Layout->id is the unique identifier that we should be using,
otherwise some layouts would never get added.
https://bugzilla.gnome.org/show_bug.cgi?id=729210
|
| |
|
| |
|
|
|
|
|
| |
gnome_wall_clock_string_for_datetime() will be used for debugging
purposes.
|
|
|
|
| |
Innocuous, but annoying.
|
|
|
|
|
|
|
| |
Since the ratio character looks visually more pleasing in time strings,
replace colons with the ratio characters whenever possible.
https://bugzilla.gnome.org/show_bug.cgi?id=726232
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
If a failed thumbnail is created because the pixbuf fail to save
(for instance if user is over quota) we should still try to rename
the temporary file that might have been created.
If not, the thumbnail will not be marked as failed, and thumbnailing
will be reattempted.
https://bugzilla.gnome.org/show_bug.cgi?id=728775
|
|
|
|
|
|
|
|
| |
This change will make the debug application listen for changes and print
them as they occur.
Conflicts:
libgnome-desktop/gnome-rr-debug.c
|
|
|
|
|
| |
On OpenBSD clock_gettime(2) is part of libc and there is not librt, so
add an according check.
|
|
|
|
|
|
| |
Fallout from 817fd136603bf92294411345b2281b6ab0d37adf
https://bugzilla.gnome.org/show_bug.cgi?id=727034
|
|
|
|
|
|
|
|
|
|
|
| |
Otherwise, the GnomeRROutputInfos representing connected but inactive
outputs (ie. without an assigned CRTC) had their rotations set to
GNOME_RR_ROTATION_NEXT, which is not a valid value. This leads to
gnome_rr_config_applicable returning FALSE, because it thinks that one
of the outputs require a rotation that is not supported by any of the
available CRTCs.
https://bugzilla.gnome.org/show_bug.cgi?id=727023
|
|
|
|
| |
I forgot to do it before pushing.
|
|
|
|
|
|
|
|
|
|
|
| |
Our cached DPMS mode value can be wrong in face of changes by
other clients we haven't seen yet, so always send the change
to mutter, and let it optimize away if needed (in practice,
it should never optimize, because the only reliable source
of DPMS state is the kernel driver, and several API calls
can reset DPMS without any change notification)
https://bugzilla.gnome.org/show_bug.cgi?id=726055
|
|
|
|
|
|
|
| |
If a variant doesn't specify language/country metadata then we should
file it under its main layout's language/country.
https://bugzilla.gnome.org/show_bug.cgi?id=711291
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This code was originally from gnome-settings-daemon but got moved here
so that we could be sure that all GnomeXkbInfo users were using the
same xkb rules file by getting it from the X server.
As we move towards wayland though, we find ourselves in a position to
specify how and where xkb data resides so we can drop this
dependency. Instead, for now, we'll just hardcode the path to the xkb
data from the xkeyboard-config XKB_DATA compile time variable.
https://bugzilla.gnome.org/show_bug.cgi?id=719942
|
|
|
|
|
|
|
|
|
| |
Some of our translations use the ratio symbol which isn't convertible
to non-UTF-8 locale encodings which causes g_date_time_format() to
fail. Work around it by replacing the ratio symbol with a plain
colon.
https://bugzilla.gnome.org/show_bug.cgi?id=722664
|
| |
|
|
|
|
|
|
|
| |
Export a method on GnomeRROutputInfo to check which rotation options are
supported.
https://bugzilla.gnome.org/show_bug.cgi?id=721672
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=719928
|
|
|
|
|
|
|
| |
This include file is no longer needed since the switch to
"locale -a" for listing locales.
https://bugzilla.gnome.org/show_bug.cgi?id=720815
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=711604
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The fact that we know about a given device doesn't mean that mutter
also knows about it nor that it has created an idle monitor object on
the bus for it.
To fix this race, instead of immediately trying to create a proxy for
the bus object, we instantiate an object manager and ask it whether
the object we want already exists and if it doesn't we wait
(indefinitely) until it shows up.
https://bugzilla.gnome.org/show_bug.cgi?id=706229
|
| |
|
|
|
|
|
|
|
| |
This shuts up a gcc warning by adding support for thumbnail directories being
unmounted or moved.
https://bugzilla.gnome.org/show_bug.cgi?id=709819
|
|
|
|
|
|
|
| |
Allows gnome-settings-daemon to read the min_backlight_step
so that it can support displays with less than 10 steps
https://bugzilla.gnome.org/show_bug.cgi?id=710380
|
|
|
|
| |
Thanks to Alexander Todorov.
|
|
|
|
|
|
|
|
| |
This prevents gdk_pixbuf_loader_close() from emitting a critical
error due to being called with a NULL loader. The loader can be NULL
if there was an error in the first read from the input stream.
https://bugzilla.gnome.org/show_bug.cgi?id=709819
|