lp:~raof/mir/maybe-less-confusing-client-visible-error

Branch merges

Propose for merging

No branches dependent on this one.

Rejected for merging into lp:mir

Daniel van Vugt: Needs Fixing on 2016-12-08

Alan Griffiths: Needs Fixing on 2016-12-07

Cemil Azizoglu (community): Abstain on 2016-11-28

Kevin DuBois (community): Abstain on 2016-11-22

Mir CI Bot: Approve (continuous-integration) on 2016-11-17

Brandon Schaefer (community): Approve on 2016-11-17

Diff: 375 lines (+36/-36)

22 files modified

include/server/mir/frontend/session.h (+2/-2)
include/test/mir/test/doubles/stub_session.h (+1/-1)
src/include/common/mir/client_visible_error.h (+3/-3)
src/include/server/mir/frontend/event_sink.h (+2/-2)
src/include/server/mir/frontend/template_protobuf_message_processor.h (+1/-1)
src/server/frontend/authorizing_display_changer.cpp (+2/-2)
src/server/frontend/event_sender.cpp (+1/-1)
src/server/frontend/event_sender.h (+1/-1)
src/server/frontend/protobuf_message_processor.cpp (+1/-1)
src/server/scene/application_session.cpp (+1/-1)
src/server/scene/application_session.h (+1/-1)
src/server/scene/global_event_sender.cpp (+1/-1)
src/server/scene/global_event_sender.h (+1/-1)
src/server/scene/mediating_display_changer.cpp (+6/-6)
tests/include/mir/test/doubles/mock_event_sink.h (+1/-1)
tests/include/mir/test/doubles/mock_scene_session.h (+1/-1)
tests/include/mir/test/doubles/null_event_sink.h (+1/-1)
tests/integration-tests/test_buffer_scheduling.cpp (+1/-1)
tests/mir_test_doubles/mock_event_sink_factory.cpp (+2/-2)
tests/mir_test_framework/stub_session.cpp (+1/-1)
tests/unit-tests/frontend/test_event_sender.cpp (+3/-3)
tests/unit-tests/scene/test_application_session.cpp (+2/-2)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

3825. By Chris Halse Rogers on 2016-11-16

Rename ClientVisibleError to StructuredClientError.

A number of people have been confused by the ClientVisibleError name, as other errors are
*also* visible to the client, just not as a structured MirError.

Try this shade of blue on the bikeshed.

3824. By Daniel van Vugt on 2016-11-16

Fix memory leak LP: #1628794, I think.

Also remove an unused (masked) member variable. Fixes: https://bugs.launchpad.net/bugs/1628794.

Approved by Cemil Azizoglu, Alan Griffiths, mir-ci-bot.

3823. By Chris Halse Rogers on 2016-11-15

Replace the mir::Server-overridable Reports with Observers.

This resolves two problems:
1) It ensures that Mir will respect the various reporting options we provide, even when the shell wants to hook into some of the reports, and
2) It matches the semantic expectations around Reports vs Observers that (most of) the Mir team have. A couple of report call-sites which are unsafe for Observers are fixed in this branch.

Approved by mir-ci-bot, Kevin DuBois, Andreas Pokorny, Cemil Azizoglu.

3822. By Brandon Schaefer on 2016-11-15

Turn off LTO for s390x builds where it causes build failures. This
unblocks the release of Mir 0.25.0 and should probably be kept in
lp:mir too.
.

Approved by Brandon Schaefer, mir-ci-bot.

3821. By Andreas Pokorny on 2016-11-14

Make FakeDisplay preserve display buffers when the displayed region stays unchanged. Fixes: https://bugs.launchpad.net/bugs/1556142.

Approved by Alan Griffiths, Chris Halse Rogers, mir-ci-bot.

3820. By Kevin DuBois on 2016-11-14

client: propose that MirBuffer is the EGLNativePixmapType for the mir platform. Include an example of buffer import to GL texture via the mir client api.

Also introduce the MIR_EXTENSION_ANDROID_EGL, with hooks for getting the EGLNativePixmapType and EGLNativeDisplayType. EGLNativeWindowType will be added after some MirRenderSurface improvements are proposed.

note: egldiamond will only import buffers on android, as the example shim subloads EGL_NATIVE_BUFFER_ANDROID. Other platforms have a sensible fallback.

Approved by Andreas Pokorny, mir-ci-bot, Chris Halse Rogers.

3819. By Alan Griffiths on 2016-11-12

Fix code to match the documentation of CoordinateTranslator - vis "It is acceptable ...[to] throw a std::runtime_error". Fixes: https://bugs.launchpad.net/bugs/1641166.

Approved by Andreas Pokorny, mir-ci-bot, Stephen M. Webb.

3818. By Daniel van Vugt on 2016-11-11

Accommodate scheduling issues on slow/loaded machines that could cause
the frame dropping test to occasionally miss a frame. Use the average
rather than worst case in order to ignore such glitches (LP: #1640366)
. Fixes: https://bugs.launchpad.net/bugs/1640366.

Approved by Kevin DuBois, Chris Halse Rogers, mir-ci-bot.

3817. By Daniel van Vugt on 2016-11-11

Enhance the ClientLatency acceptance tests to measure latency with
fractional precision.

This is required to test framedropping properly, which should always
achieve closer to zero than one whole frame of latency. In the new
Vulkan terminology, this is called Mailbox mode.

This is an important regression test not only for maintaining low
latency but also avoiding bugs like LP: #1379685 and LP: #1396006. Fixes: https://bugs.launchpad.net/bugs/1640366.

Approved by mir-ci-bot.

3816. By Kevin DuBois on 2016-11-11

client: add a mechanism for adding extensions to the client API.

Approved by Andreas Pokorny, mir-ci-bot, Chris Halse Rogers.