This fixes a problem identified by devin.ai: Because of the operator
precedence the previous code caused `memmove` to read
`sizeof(commit_queue_item) - 1` bytes past the end of `commit_queue`.
With this change `memmove` now reads the intended `MAX_QUEUE_SIZE - 1`
`commit_queue_items`.
Test-bot: skip
This fixes a problem with the context after typing Backspace when
using Wayland. Wayland uses double-buffering for the text, so we always
have to commit after making changes.
Fixes: #15676
This change adds the verification of the options to the baseline tests
and fixes persisting the options. Previously we treated the UTF-16 key
and value as UTF-8 so that only the first character got persisted.
Test-bot: skip
This change suppresses the output of ending the processes started for
the test run. The output doesn't help much but clutters the overall
output and shows up as warnings on TC.
Previously the builder scripts defined a readonly `VERSION` environment
variable for the Keyman version. That caused problems when another
(external) script tried to define a `VERSION` variable. We encountered
this problem when trying to move the TC build steps of a configuration
into a single script (#13399) when we tried to source `~/.nvm/nvm.sh`.
This change uses a Keyman specific prefix for the version variables and
renames `VERSION` → `KEYMAN_VERSION` etc. Unfortunately these variables
are used in a lot of places, so this turned out to be a bit of a yak
shave.
Test-bot: skip
fix(linux): allow unhandled keys to pass through to compliant apps
This change fixes a bug introduced in #13372. Before that engine.c determined if keys are handled or not. For keys that we didn't handle we returned FALSE. However, that didn't allow the Core to adjust the context where necessary, so #13372 moved that to Core. However, this caused ibus_keyman_engine_process_key_event to return TRUE even if we didn't handle the key. Instead we called ibus_engine_forward_key_event with the non-handled key. This worked in most applications, but not in the Text Editor.
This change now does no longer call ibus_engine_forward_key_event for compliant apps but instead returns FALSE from ibus_keyman_engine_process_key_event for unhandled keys, allowing the app to see and act on the key event.
An 'unhandled' key here is a key that doesn't match a rule in the keyboard and isn't a character key. Core returns a QIT_EMIT_KEYSTROKE action for that key.
Fixes: #13590
This change fixes a bug introduced in #13372. Before that `engine.c`
determined if keys are handled or not. For keys that we didn't handle
we returned `FALSE`. However, that didn't allow the Core to adjust the
context where necessary, so #13372 moved that to Core. However, this
caused `ibus_keyman_engine_process_key_event` to return `TRUE` even if
we didn't handle the key. Instead we called `ibus_engine_forward_key_event`
with the non-handled key. This worked in most applications, but not
in the Text Editor.
This change now does no longer call `ibus_engine_forward_key_event` for
compliant apps but instead returns `FALSE` from
`ibus_keyman_engine_process_key_event` for unhandled keys, allowing the
app to see and act on the key event.
Fixes: #13590
The previous code resulted in warnings showing up:
`(ibus-engine-keyman:4276): GLib-CRITICAL **: 14:31:16.030: g_hash_table_lookup: assertion 'hash_table != NULL' failed`.
This should work around the warning.
Several keys that are defined in `winuser.h` were added after Keyman
defined its vkeys and so are not documented in the Keyman docs. This
change adds underscores before and after to mark them as being internal.
It's still useful to have them in the source code because it makes the
mapping between Linux keys and Keyman vkeys easier.
Addresses code review comment.
This change tries to map all keys to corresponding Core VKeys so that
Core sees them and can reset the context if necessary. Previously we
had 0 for most of the keys, which meant that we didn't send them to
Core but instead returned right away. While this doesn't directly fix
the failing user tests described in #12968, it fixes the original tests
in #11172.
Closes: #12968
For some reason when using a LDML keyboard we ended up emitting a
keystroke even though Core returned `FALSE for `emitting_keystroke`.
I wasn't able to track down where this happened, but this change seems
like an easy way to fix it.
This showed up when debugging why the changes in #13372 didn't work for
non-compliant apps.
The new name `initialize_queue_items` makes it clearer that we initialize
some items in the queue and not the entire queue (which might still
contain items that we do want to keep).
This change contains various changes and refactorings which together
allows the integration tests to pass again.
One of the important changs is `DbusTestHelper` which listens to a DBus
signal for a sentinel message and then simulates the sentinel key press.
Getting this to work also requires processing sd-bus messages which hooks
into the glib event loop and is implemented in `EventSource`.
Fixes: #13204
This change fixes#13171 where we failed to output after the first keypress
because the keyman system service had to be started. This change introduces
a new method `Ping` that we call to force keyman-system-service to be
started.
Fixes: #13171
This change rename the test files for Linux according to the discussion
at the Keyman conference in November 2024.
Python is pretty opinionated about the naming of files, so we have to
use `*_tests.py` as test filename instead of the usual `*.tests.py`.
This renames the existing methods that read and write keyboard options
so that it's visible from the name that they read/write keyboard options.
This is in preparation of new methods that read and write more general
options.
If we process keys that have the (Super/Meta/Windows) IBUS_MOD4_MASK
modifier set we ignore the key. Core doesn't handle the Meta modifier
state and so would treat it as regular key. This fixes switching keyboards
in Ubuntu 23.10/24.04 on Wayland.
Fixes#10476.
