~andreas-pokorny/mir/compose-sequences-in-xkb-keymapper : changes

~andreas-pokorny/mir/compose-sequences-in-xkb-keymapper » Changes from revision 3778

From Revision 3778 to 3759

Rev	Summary	Authors	Date
3778	Make sure that key mapping and composing states stay stable ... Make sure that key mapping and composing states stay stable even when the hash container has to enlarge the hash array	Andreas Pokorny	7 years ago
3777	the invalid argument is an invalid_argument the invalid argument is an invalid_argument	Andreas Pokorny	7 years ago
3776	Improve error handling as we do have more use of options and... Improve error handling as we do have more use of options and variants, that may fail	Andreas Pokorny	7 years ago
3775	xkbcommon >= 0.5 xkbcommon >= 0.5	Andreas Pokorny	7 years ago
3774	cleanups cleanups	Andreas Pokorny	7 years ago
3773	Dead key key compose sequences in mirserver and mirclient Dead key key compose sequences in mirserver and mirclient This implementation of key compose sequences keeps the stream of scan codes intact. So clients still get the full stream - only the scan codes are removed. Hence a client with a different surface keymap could still map key presses unaffected. This is intended to be used inside qtmir to avoid qts default behavior of converting the compose sequence result into input method events, and thereby filtering out key events.	Andreas Pokorny	7 years ago
3772	Add mirtest-dev boost dependencies for xenial to debian/cont... Add mirtest-dev boost dependencies for xenial to debian/control. (LP: #1633537). Fixes: https://bugs.launchpad.net/bugs/1633537. Approved by mir-ci-bot, Cemil Azizoglu.	Alan Griffiths	7 years ago
3771	Add mirtest-dev boost dependencies to debian/control. Fixes:... Add mirtest-dev boost dependencies to debian/control. Fixes: https://bugs.launchpad.net/bugs/1633537. Approved by mir-ci-bot, Brandon Schaefer, Andreas Pokorny.	Alan Griffiths	7 years ago
3770	Switch to using CSS cursor names, as found in web design and... Switch to using CSS cursor names, as found in web design and GTK3. These names seem to be the most modern, most complete and least ambiguous of the naming schemes out there. This fixes LP: #1632617 and LP: #1626924, as well as providing a solid foundation for fixing other cursor bugs like LP: #1632614 and LP: #1388987 in future. Fixes: https://bugs.launchpad.net/bugs/1626924, https://bugs.launchpad.net/bugs/1632617. Approved by mir-ci-bot, Cemil Azizoglu, Alexandros Frantzis, Alan Griffiths.	Daniel van Vugt	7 years ago
3769	tests: Ensure client surface is in a steady state in tests u... tests: Ensure client surface is in a steady state in tests using a nested setup (LP: #1523621). Fixes: https://bugs.launchpad.net/bugs/1523621. Approved by Alan Griffiths, Cemil Azizoglu, mir-ci-bot.	Alexandros Frantzis	7 years ago
3768	Reduce input event exposure in key repeat dispatcher Reduce input event exposure in key repeat dispatcher Do not access input event types directly, and do not create key events on stack. . Approved by mir-ci-bot, Brandon Schaefer, Cemil Azizoglu, Chris Halse Rogers.	Andreas Pokorny	7 years ago
3767	Regenerate the "public" part of src/common/symbols.map, labe... Regenerate the "public" part of src/common/symbols.map, label the "private" and deduplicate a few symbols. Approved by Cemil Azizoglu, mir-ci-bot.	Alan Griffiths	7 years ago
3766	Add a POSIX implementation of C++14's SharedMutex concept. Add a POSIX implementation of C++14's SharedMutex concept. The reason to use this over std::shared_timed_mutex is that POSIX makes a guarantee that read locks are recursive, wheras this is undefined behaviour for std::shared_timed_mutex. Approved by mir-ci-bot, Cemil Azizoglu, Alan Griffiths, Alexandros Frantzis.	Christopher James Ha...	7 years ago
3765	Rework Validator to not need direct event structure access Rework Validator to not need direct event structure access The rework also cleans up some boilerplate code an makes the order in which missing contact state changes are emitted more sorted than before - hence the small change in the test case. The missing ups and downs are applied in the order they occurred in the first place. Approved by Cemil Azizoglu, Brandon Schaefer, mir-ci-bot.	Andreas Pokorny	7 years ago
3764	Move the InputDeviceState into a shared region we can then u... Move the InputDeviceState into a shared region we can then use for both event_builder and the class InputDevice. This is needed for capnproto changes due to capnproto needing to know the full size of the number of InputDeviceStates so adding one at a time doesnt work. Approved by mir-ci-bot, Cemil Azizoglu, Alan Griffiths.	Brandon Schaefer	7 years ago
3763	Move headers incorrectly moved to libmircore-dev back to lib... Move headers incorrectly moved to libmircore-dev back to libmircommon-dev. Approved by Andreas Pokorny, mir-ci-bot.	Alan Griffiths	7 years ago
3762	Stop publishing the mircommon headers through mirclient-dev. Stop publishing the mircommon headers through mirclient-dev. Approved by Chris Halse Rogers, Cemil Azizoglu, mir-ci-bot.	Alan Griffiths	7 years ago
3761	Introduce libmircore.so: an ABI-stable library of Mir's "bas... Introduce libmircore.so: an ABI-stable library of Mir's "basic vocabulary". Approved by mir-ci-bot, Brandon Schaefer, Andreas Pokorny, Kevin DuBois.	Alan Griffiths	7 years ago
3760	Add an option to disable passthrough optimization in nested ... Add an option to disable passthrough optimization in nested servers. Option is defaulted to passthrough-on. The option is useful for debugging and performance inspection. Approved by mir-ci-bot, Daniel van Vugt, Alexandros Frantzis, Brandon Schaefer.	Kevin DuBois	7 years ago
3759	android: disable overlays when a swapinterval 0 client is pa... android: disable overlays when a swapinterval 0 client is part of the RenderableList. This won't affect the overlay percentage for usc/u8, as u8 runs swapinterval 1. mg::Renderable providing swapinterval information is useful to android to fix this bug, as well as for nested-passthrough, where nested-passthrough has to coordinate the passthrough surface with the host server, so the host server doesn't end up grabbing all the buffers from the passthrough surface. platform abi number was already increased for the 0.25 series. Fixes: LP: #1369763. Fixes: https://bugs.launchpad.net/bugs/1369763. Approved by Daniel van Vugt, Brandon Schaefer, mir-ci-bot, Cemil Azizoglu.	Kevin DuBois	7 years ago

Older »