| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
| |
Fixes https://gitlab.gnome.org/GNOME/at-spi2-atk/issues/2
|
| |
|
|\
| |
| |
| |
| | |
tests: fix child_pid reference
See merge request GNOME/at-spi2-atk!13
|
|/
|
|
|
|
| |
180f91479e27 ("tests: Replace common variable with usual external
variable") fixed the variable declaration, but there was no definition
either. This adds it.
|
|\
| |
| |
| |
| | |
cache-adaptor: Fix crash when no toolkit name is set
See merge request GNOME/at-spi2-atk!12
|
| |
| |
| |
| |
| |
| | |
atk_get_toolkit_name may return NULL when e.g. running the pyatspi2 test
suite. Callers of get_toolkit_name expect a non-NULL value (passed to
strcmp), so return a default string.
|
|\ \
| | |
| | |
| | |
| | | |
tests: Replace common variable with usual external variable
See merge request GNOME/at-spi2-atk!11
|
| |/ |
|
|\ \
| | |
| | |
| | |
| | | |
Fix typo
See merge request GNOME/at-spi2-atk!10
|
| |/ |
|
|\ \
| |/
|/|
| |
| | |
tests: Fix running atk-test
See merge request GNOME/at-spi2-atk!9
|
| |
| |
| |
| |
| | |
Some archs might round them differently from xml and from C, leading to
float comparison failure, e.g. on i386.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This allows to reduce the testsuite completion time from almost 3 minutes to
about 3s.
This however showed that we were not waiting for the application to
terminate. Add terminate_app() which ensures this, to be used instead of
only killing the application.
|
|/
|
|
|
|
|
|
| |
- Actually add it to the meson tests.
- Let run_app find app-test when run from the build root by ninja by
making it use the build dir.
- It takes almost three minutes due to waiting for 1s for application
startup, give it 5 minutes to run.
|
| |
|
|\
| |
| |
| |
| | |
Make sure returned values are initialized
See merge request GNOME/at-spi2-atk!8
|
| |
| |
| |
| |
| |
| |
| |
| | |
Some methods do not have a way to notify that they have failed.
They should thus make sure that they set some value, rather than let them
uninitialized and thus random.
As newly documented by https://gitlab.gnome.org/GNOME/atk/merge_requests/22
|
|\ \
| |/
|/|
| |
| |
| |
| | |
Clear registration_pending during registration
Closes at-spi2-core#16
See merge request GNOME/at-spi2-atk!7
|
|/
|
|
|
|
|
| |
Otherwise switch_main_context will keep trying to remove it and schedule
yet another one.
Fixes https://gitlab.gnome.org/GNOME/at-spi2-core/issues/16
|
|
|
|
|
|
|
| |
Commit f603c8 broke java-atk-wrapper, since the default main loop context
isn't used there.
Fixes https://gitlab.gnome.org/GNOME/at-spi2-core/issues/16
|
|
|
|
|
|
|
|
|
| |
A python application, for instance, may initialize gtk but not enter the
main loop. It would, thus, register with the AT-SPI registry but not be
responsive to queries. Putting registration into an idle callback ensures
that, if a program never enters the main loop, then it won't register.
Fixes https://gitlab.gnome.org/GNOME/at-spi2-core/issues/16
|
| |
|
|
|
|
| |
Fixes: https://gitlab.gnome.org/GNOME/at-spi2-atk/issues/8
|
| |
|
|\
| |
| |
| |
| | |
Add support for ROLE_CONTENT_DELETION and ROLE_CONTENT_INSERTION
See merge request GNOME/at-spi2-atk!6
|
|/ |
|
| |
|
|\
| |
| |
| |
| | |
Add AccessibleId property
See merge request GNOME/at-spi2-atk!5
|
|/ |
|
|\
| |
| |
| |
| | |
Add atk_bridge_set_event_context
See merge request GNOME/at-spi2-atk!3
|
| |
| |
| |
| |
| |
| |
| | |
to allow java-atk-wrapper to control in which thread events are
delivered.
See https://bugzilla.gnome.org/show_bug.cgi?id=760058
|
|\ \
| | |
| | |
| | |
| | | |
atk-adaptor: document atk_bridge_adaptor_init return messages
See merge request GNOME/at-spi2-atk!4
|
|/ / |
|
|/ |
|
| |
|
|
|
|
| |
Fixes https://gitlab.gnome.org/GNOME/at-spi2-atk/issues/5
|
| |
|
|\
| |
| |
| |
| | |
Add ScrollSubstringTo and ScrollSubstringToPoint text interfaces
See merge request GNOME/at-spi2-atk!1
|
|/ |
|
| |
|
| |
|
|
|
|
|
| |
The AtkComponent scrolling methods have been introduced in ATK 2.29.2,
so we need to require at least that version.
|
|
|
|
| |
And bring both a bit more up to date.
|
|
|
|
|
| |
We should always use `#include "config.h"`, as the header is local to
the project, not part of the system include paths.
|
|
|
|
|
|
| |
All major build systems and distributors downstream know how to handle
Meson, these days, so there's no need to provide an Autotools
compatibility wrapper.
|
|
|
|
|
| |
Instead of using a template, we can ask Meson to generate the file for
us from the various build targets.
|
|
|
|
| |
We never even assigned its return value.
|
|
|
|
|
| |
The tests are still meant to be used by hand, not as part of a test
suite.
|
|
|
|
|
|
|
|
| |
We know we're using deprecated API, so there's no real point in keeping
the warnings.
The deprecation warnings are useful during development cycles, so we can
port away from deprecated API.
|
|
|
|
| |
Leftovers from previous iterations.
|