Per the Apple documentation, when 0 is returned from
IOSurfaceGetPixelFormat, the IOSurface is invalid. Prior to [1] however,
any invalid surfaces were (likely accidentally) treated as 32BGRA.
And as it turns out, any IOSurface received from Syphon does not have
the pixel format set, thus returning 0. Because the pixel format happens
to always be 32BGRA, this was never caught. [1] then slightly refactored
the code, fixing the bug not realizing it ever existed. This lead to
Syphon no longer working.
Treating 0 as 32BGRA appears to be the only way to fix this situation
for now and restore prior behavior. Ideally in the future Syphon
provides valid IOSurfaces. An upstream fix to Syphon-Framework is
already submitted, however even if merged, servers would need to update
to that fixed version before things work again. Therefore, this
workaround is needed on the OBS side.
However, the plan for now is for the Metal renderer to not get this
relaxation. It never had it (so Syphon never worked there, making this
not a regression) and it's marked as experimental. Hopefully, by the
time it's stable and/or default, Syphon will be fixed and servers have
been updated.
[1]: 53ad05db9f
By explicitly setting the OAuth dialog as the parent of the CEF widget,
it will automatically by cleaned up as soon as the dialog is closed
itself.
This also detaches the destruction of the widget from the browser
cleanup code.
The commit adds a log line to inform the user that deduplication is
being applied. The info is displayed whenever deduplication first
occurs.
Signed-off-by: pkv <pkv@obsproject.com>
When an 'Audio Capture Source' device is also used for monitoring, the
deduplication logic is applied: all monitored sources are silenced.
But this should not silence the 'Audio Capture Source' if for some
reason, it is being monitored (the user will get echoes, but hey, it's
their choice). So we exclude the 'Audio Capture Source' from the
silenced sources.
Signed-off-by: pkv <pkv@obsproject.com>
This removes the need to provide an undocumented launch argument to
enable custom RTMP service configurations to use multitrack encoding
and also provide a custom configuration.
The default of 2500 kbps was chosen 10 years ago and times have changed.
Logs and forums posts show that many users of OBS for recording don't
change their bitrate and end up with a 2.5 mbps recording which looks
terrible.
Now that service bitrate enforcement exists, this will be automatically
capped to the maximum bitrate for streaming services, so the only time
this should result in a problem is if the user's upload speed is the
limiting factor, hopefully rarer these days with increasing internet
speeds.
The previous attempt to clarify an out of memory exception/crash
resulted in the compiler optimizing os_breakpoint() and os_oom() into
the same result, which meant that crash stacks in the wild were still
not specific enough to be helpful. Forcefully differentiating the
functions in Release configuration by having os_breakpoint() only call
__debugbreak() and having os_oom() call RaiseException() should give us
clearer crash stacks.
Amends 94a736f179.
There is a Qt bug introduced around Qt 6.8.3 that causes the application
to hang when double clicking the sources list and the Windows setting
'Snap mouse to default button in dialog boxes' is enabled. Work around
this by using a timer if the setting is enabled.
The important point here is that "data" from the callback isn't defined
by libobs, but rather by the API consumer.
Additionally, it's not entirely clear from the documentation whether the
callback would still be called for button of type OBS_BUTTON_URL. There
should be no negatives to always calling it and the OBS Studio frontend
already behaves that way, so let's document that in libobs.
It's important to differentiate whether a plugin fails to open (dl_open
returning NULL, being compiled for a newer libobs, etc) or whether it
opened correctly, but failed to initialize (returned NULL during
obs_module_load()).
With that, we can also create disabled modules for modules that fail to
open, which in the future should enable better user communication in the
UI (besides the module_load_failure_info).
This code is returned if os_dlopen returns NULL. This can happen for
multiple reasons, not just because the file can't be found. Since [1]
other causes are getting more common, but this could also happen before.
[1]: 62429135ba
While the Metal renderer is displayed as "Metal (Experimental)", the
"Experimental" part doesn't need to be stored to the config. Otherwise,
anyone having selected the Metal renderer will have it get unselected
once the "Experimental" part is removed, and presumably the people who
have it selected while it's experimental would also be the people who
would want it once it's stable.
The current version of swift-format uses features that are only
supported in more recent Swift compiler versions than the one shipped
with Xcode 16.1.0.
Partial revert of a0eae6f33c.
Partial revert of 23c3ad4d02.
Partial revert of 97b34ebb76.
Keep all of the get_plugin_info stuff, remove the Qt5 checks.
Ryan Foster
Sebastian BeckmannandRyan Foster
PatTheMav
hongqingwanandRyan Foster
pkvandRyan Foster
Dennis SädtlerandRyan Foster
nixflorpandRyan Foster
tytan652andRyan Foster
Richard StanwayandRyan Foster
Joel BethkeandRyan Foster
Service CheckerandRyan Foster
Norihiro KamaeandRyan Foster
PenwywernandRyan Foster
Warchamp7andRyan Foster
derrodandRyan Foster
Matt GajownikandRyan Foster
Amin.MasterkinGandRyan Foster