Alexandre Julliard pushed to branch master at wine / wine
Commits:
332f891f by Paul Gofman at 2025-10-15T11:43:24+02:00
windows.media.speech: Add stub IRandomAccessStream interface to SpeechSynthesisStream class.
- - - - -
30452cc6 by Paul Gofman at 2025-10-15T11:43:26+02:00
windows.media.speech: Add stub IInputStream interface to SpeechSynthesisStream class.
- - - - -
84cd9757 by Paul Gofman at 2025-10-15T11:43:28+02:00
windows.media.speech: Add semi-stub for synthesis_stream_input_ReadAsync().
- - - - -
09622456 by Paul Gofman at 2025-10-15T11:43:29+02:00
windows.media.speech: Return semi-stub IAsyncInfo from async_with_progress_uint32_QueryInterface().
- - - - -
3 changed files:
- dlls/windows.media.speech/private.h
- dlls/windows.media.speech/synthesizer.c
- dlls/windows.media.speech/tests/speech.c
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/compare/29065fdc8994ddc16e2bed29ef9d5…
--
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/compare/29065fdc8994ddc16e2bed29ef9d5…
You're receiving this email because of your account on gitlab.winehq.org.
Alexandre Julliard pushed to branch master at wine / wine
Commits:
79b8ed9f by Giovanni Mascellani at 2025-10-15T11:42:43+02:00
dsound: Remove dead lead-in logic.
This is effectively dead code, because if the buffer's state is
STARTING it gets changed to PLAYING just before calling
DSOUND_MixOne(). It has been like that since
6e8c8c28ebbdc15292612b7b6657e44ce8c5040f (17 years ago), mere
months after the lead-in logic was introduced in
36e90546298e8c096926330dd75f4e373a415171.
I don't know whether that code does something useful. It was
introduced in a monster 390+ 310- commit whose commit message
doesn't say much, and the comment near the code only mentions
"more fluid pointer advancement", which I have no idea of
what it means. But the fact that it was effectively dead for
95% of its existence and nobody every complained seems to
suggest that we can get rid of it.
- - - - -
ee73c569 by Giovanni Mascellani at 2025-10-15T11:42:45+02:00
dsound: Simplify a condition in DSOUND_MixInBuffer().
Similarly to the previous commit, the state can never be STARTING
at that place.
- - - - -
3 changed files:
- dlls/dsound/buffer.c
- dlls/dsound/dsound_private.h
- dlls/dsound/mixer.c
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/compare/93a3d8bf9f15b9199946f6e948436…
--
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/compare/93a3d8bf9f15b9199946f6e948436…
You're receiving this email because of your account on gitlab.winehq.org.
Alexandre Julliard pushed to branch master at wine / wine
Commits:
93a3d8bf by Yuxuan Shui at 2025-10-15T11:42:30+02:00
iphlpapi: Give GetBestRoute2 a SOCKETADDR_INET.
In this call chain: GetBestInterface -> GetBestInterfaceEx -> GetBestRoute2, we cast a socketaddr_in
* (16 bytes) is to SOCKETADDR_INET * (28 bytes) before passing it to GetBestRoute2, but
GetBestRoute2 will access all of the 28 bytes, leading to out-of-bound access. So we create a local
SOCKETADDR_INET in GetBestInterfaceEx.
Found by ASan.
- - - - -
1 changed file:
- dlls/iphlpapi/iphlpapi_main.c
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/commit/93a3d8bf9f15b9199946f6e948436b…
--
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/commit/93a3d8bf9f15b9199946f6e948436b…
You're receiving this email because of your account on gitlab.winehq.org.
Alexandre Julliard pushed to branch master at wine / wine
Commits:
5f793f78 by Giovanni Mascellani at 2025-10-15T11:42:03+02:00
mmdevapi/tests: Iterate independently on sampling rates, channel counts and sample formats when rendering.
- - - - -
5ec27192 by Giovanni Mascellani at 2025-10-15T11:42:03+02:00
mmdevapi/tests: Tweak the lists of audio format to test.
Trying to balance between coverage and running time. Especially on
macOS it takes a lot of time to initialize and deinitialize the
audio system over and over.
- - - - -
5b9edf29 by Giovanni Mascellani at 2025-10-15T11:42:07+02:00
mmdevapi/tests: Simplify checking IsFormatSupported() result when rendering.
- - - - -
4ecfb4de by Giovanni Mascellani at 2025-10-15T11:42:07+02:00
mmdevapi/tests: Check that Initialize() matches IsFormatSupported() when rendering.
- - - - -
f899efc3 by Giovanni Mascellani at 2025-10-15T11:42:07+02:00
mmdevapi/tests: Test flag AUDCLNT_STREAMFLAGS_RATEADJUST when rendering.
- - - - -
8171dd5e by Giovanni Mascellani at 2025-10-15T11:42:07+02:00
mmdevapi/tests: Test flag AUDCLNT_STREAMFLAGS_AUTOCONVERTPCM when rendering.
- - - - -
716a4af8 by Giovanni Mascellani at 2025-10-15T11:42:07+02:00
mmdevapi/tests: Test extensible wave formats when rendering.
This adds a few more cases in which initializing in exclusive
mode can fail even when IsFormatSupported() returned positively.
To avoid making the broken() conditions overly complicated I
remove the side conditions, which are probably not very useful
and depend on the driver anyway.
Also, testing configurations with more than two channels somehow
alters the state of the device in a way that is not restored
even after releasing the IMMDevice object and creating it again.
Specifically, GetChannelVolume() will return more then two
channels, even if the audio client is initialized with just two
channels. To avoid those failures the format tests are moved
to the end.
- - - - -
1 changed file:
- dlls/mmdevapi/tests/render.c
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/compare/120db075b5e8863c1ded907e38a88…
--
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/compare/120db075b5e8863c1ded907e38a88…
You're receiving this email because of your account on gitlab.winehq.org.
Alexandre Julliard pushed to branch master at wine / wine
Commits:
a4fbcc45 by Yuxuan Shui at 2025-10-15T11:39:08+02:00
windows.devices.enumeration: Fix use-after-realloc in devpropcompkeys_append_names.
- - - - -
6bf9d368 by Yuxuan Shui at 2025-10-15T11:39:10+02:00
windows.devices.enumeration: Don't leave dangling pointers if devpropcompkeys_append_names fails.
If devpropcompkeys_append_names fails after the realloc, *ret_keys will contain a dangling pointer.
But both of its callers assume it's safe to call free on *ret_keys when
devpropcompkeys_append_names fails, resulting in double freeing.
- - - - -
1 changed file:
- dlls/windows.devices.enumeration/main.c
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/compare/6bb444a3c82d0f5e9ff3190352e63…
--
View it on GitLab: https://gitlab.winehq.org/wine/wine/-/compare/6bb444a3c82d0f5e9ff3190352e63…
You're receiving this email because of your account on gitlab.winehq.org.