~raof/mir/anr-timeouts : changes

~raof/mir/anr-timeouts » Changes from revision 3688

From Revision 3679 to 3660

Rev	Summary	Authors	Date
3679	mirserver-dev headers include uuid.h so make sure our consum... mirserver-dev headers include uuid.h so make sure our consumers have that at configure time instead of them failing during build. (LP: #1617435). Fixes: https://bugs.launchpad.net/bugs/1617435. Approved by Alan Griffiths, Kevin DuBois, mir-ci-bot.	Daniel van Vugt	7 years ago
3678	Add missing build-dep for libepoxy-dev (LP: #1617256). Fixes... Add missing build-dep for libepoxy-dev (LP: #1617256). Fixes: https://bugs.launchpad.net/bugs/1617256. Approved by Alan Griffiths, Kevin DuBois, mir-ci-bot.	Daniel van Vugt	7 years ago
3677	tests: remove the frig in server.cpp that would override som... tests: remove the frig in server.cpp that would override some internal classes to avoid GL calls when running with our test modules. We can just override the_display_buffer_compositor_factory to avoid making GL calls when we're running with a stubbed platform. This became a problem when I was trying to make the nested server into its own platform .so, as detecting based solely on platform library name didn't work anymore, especially in light of the NestedServer* acceptance tests. . Approved by mir-ci-bot, Daniel van Vugt, Chris Halse Rogers.	Kevin DuBois	7 years ago
3676	Fix a bunch of races in DisplayConfigurationTest. Fix a bunch of races in DisplayConfigurationTest. FakeDisplay::wait_for_configuration_change(), as its name suggests, waits until the configuration change handler has run. Unfortunately, MediatingDisplayChanger's configuration change handler just schedules a task on the main loop, so this is insufficient to ensure that every component of the system agrees that the new configuration is active. Fixes: https://bugs.launchpad.net/mir/+bug/1563210. Fixes: https://bugs.launchpad.net/bugs/1563210. Approved by mir-ci-bot, Kevin DuBois, Alexandros Frantzis, Daniel van Vugt.	Christopher James Ha...	7 years ago
3675	Implement create_egl_native_window() in a simpler way for me... Implement create_egl_native_window() in a simpler way for mesa. Now both android and mesa platform implementations look alike. . Approved by Daniel van Vugt, Chris Halse Rogers, mir-ci-bot.	Cemil Azizoglu	7 years ago
3674	graphics: Move mg::Buffer::stride to mrs::PixelSource. mrs::... graphics: Move mg::Buffer::stride to mrs::PixelSource. mrs::PixelSource users would be the only users concerned with needing to know the stride of the buffer. Approved by mir-ci-bot, Chris Halse Rogers.	Kevin DuBois	7 years ago
3673	Remove FreeSync TODO comments. It seems we don't have to do ... Remove FreeSync TODO comments. It seems we don't have to do anything because adaptive refresh rate displays just report their maximum refresh rate in the display config. And that's all we need them to do. . Approved by mir-ci-bot, Alexandros Frantzis.	Daniel van Vugt	7 years ago
3672	Use compiler feature detection to disable false-positive -We... Use compiler feature detection to disable false-positive -Werrors. clang's -Wall -Wextra includes -Wmismatched-tags, which we don't care to (or can't) fix. Rather than trying to match against the compiler name, just check if these warnings are accepted by the compiler. This fixes the build when using compilers that support these but don't match clang, such as Clang's scan-build static analyser.	Christopher James Ha...	7 years ago
3671	When doing --gtest_repeat=2 on a few tests they would fail d... When doing --gtest_repeat=2 on a few tests they would fail due to a static variables.	Brandon Schaefer	7 years ago
3670	Disable the offending test till a proper fix is found (LP: #... Disable the offending test till a proper fix is found (LP: #1586382) This should unblock CI fully.	Daniel van Vugt	7 years ago
3669	Paranoia for UnresponsiveClient.does_not_hang_server. Paranoia for UnresponsiveClient.does_not_hang_server. This resolves a failure that I very infrequently see when running under valgrind under load.	Christopher James Ha...	7 years ago
3668	We use "C" linkage to avoid name-mangling some functions tha... We use "C" linkage to avoid name-mangling some functions that are not intended not for C compatibility. We don't want downstream projects seeing a warning from clang based tools for us doing this intentionally. (LP: #1615587)	Alan Griffiths	7 years ago
3667	Don't use namespace aliases in public headers (LP: #1614983) Don't use namespace aliases in public headers (LP: #1614983)	Daniel van Vugt	7 years ago
3666	Add mir_output_get_{current,preferred}_mode_index API. Add mir_output_get_{current,preferred}_mode_index API. This is trivial to retrieve in the client library, and I've found that I've been smattering ad-hoc, worse, implementations of it around in my client code. So just expose it in the client API and make everyone's lives easier.	Christopher James Ha...	7 years ago
3665	Simplify Rectangle logic, so that it's easier to follow. Simplify Rectangle logic, so that it's easier to follow.	Daniel van Vugt	7 years ago
3664	changelog: Merge final 0.24.0 changelog from Ubuntu and add ... changelog: Merge final 0.24.0 changelog from Ubuntu and add an entry for 0.25.0	Daniel van Vugt	7 years ago
3663	Remove one line of dead code. Remove one line of dead code.	Daniel van Vugt	7 years ago
3662	backout rev 3650 so that newer PowerVR devices (like mx4) do... backout rev 3650 so that newer PowerVR devices (like mx4) don't try to close the old FB module (which segfaults when trying to close). Seeing as how its better to just load the fb module for devices that actually need this, We should have a quirk for devices that need this (probably just omap devices) fixes: LP: #1615703	Kevin DuBois	7 years ago
3661	android: create HwcBlankingControl as HWC configuration on H... android: create HwcBlankingControl as HWC configuration on HWC 1.0 This changes screen initialization on a device using HWC 1.0 from the FB HAL way to the HWC way. This makes vsync works, as previously vsync is not enabled in FbControl. This fix, together with the hack from lp:~peat-new/mir/hack-eglswapbuffer-call-fbpost , makes graphics with HWComposer works on Samsung Galaxy Tab 2 7.0. So, to properly support this device, the hack must be verified if it's the correct way to do. This is beyond my knowledge and requires a developer to have a look at it.	Ratchanan Srirattana...	7 years ago
3660	platforms: remove ifdefs around mg::NativeBuffer to make it ... platforms: remove ifdefs around mg::NativeBuffer to make it simply a forward-declared class for server code. Platforms can define what mg::NativeBuffer is to whatever works best. (eg, Android defines it to be an interface that makes it easy to convert to different android types and manage the buffer fences). Android defines it to be an interface that makes it easy to convert to different android types and manage the buffer fences. Mesa was hacking around the POD type it was using with GBMNativeBuffer to get the gbm_bo. GBMNativeBuffer is just mg::NativeBuffer for that platform now. Eglstream could use ShmFile as its native type (although I didn't make that change in this MP) Nested will want a mg::NativeBuffer that gives access to its underlying MirBuffer. Approved by mir-ci-bot, Chris Halse Rogers, Alexandros Frantzis.	Kevin DuBois	7 years ago

Older »