Instead of a checkbox in the recording group that reveals the replay
buffer group when pressed, the replay buffer group now has its own
toggle that enables or disables the replay buffer.
When running an Output Timer multiple times, multiple signal-slot
connections would be made with the corresponding output stop event
functions. This would cause later Output Timers to terminate
unexpectedly early, due to the stop functions having been called
repeatedly on previous Output Timer runs while outputs were stopping.
There may still be an underlying bug with calling stop on outputs
repeatedly, but this change at least ensures that Output Timers only
call their stop functions once.
Group resize is not deferred until adjustments are completed by a
mouseReleaseEvent in certain cases, resulting in unexpected movement of
sources that are part of a group as they are resized or stretched
beyond group bounds.
The bug can be described/reproduced as follow:
1. The user selects within a group in the sources dock
2. The user moves their mouse over the selected source such that the
cursor would change.
- This results in a cursor update and a call to GetStrechHandle()
- as the source is part of a group, stretchGroup will be set to the
group of the source
3. The user clicks on the canvas without touching a stretch/rotation
handle
- A mouseReleaseEvent is fired
- obs_sceneitem_defer_group_resize_end is called on the group as
stretchGroup is still defined from earlier
- the defer_group_resize member of the group will become negative
- The deferal check in resize_group (obs-scene.c) will always pass
now
4. When scaling or rotating the source close to the group bounds, the
group bounds will now dynamically update causing the source in to fly
off the canvas.
Resolves: https://github.com/obsproject/obs-studio/issues/9754
Loading these PNGs in libpng 1.6+ results in the warning:
libpng warning: iCCP: known incorrect sRGB profile
This is probably caused by saving an ICC Profile with the program used
to make the image, and that ICC Profile is considered invalid by libpng.
Removing the iCCP data resolves this.
Didn't realize OBS could output both program and preview views at the
same time with multiple devices.
Also remove render callbacks earlier on stop to avoid use-after-free.
Fixes an issue where if no filters had previously been copied and the
user copies a filter from either the scene tree or the audio mixer, the
"Paste Filters" action would not be available until the Edit menu
reloads (e.g., by selecting another source in the source tree).
The YouTubeAppDock uses its own cookie manager and thus requires a
running CEF instance before creating the dock. Unfortunately creation
of the dock itself and launching the associated browser instance are
coupled in the code.
The UI code to restore browser dock states runs _after_ this code and
unfortunately this is also the only way to ensure that if the user
has closed the YouTubeAppDock before that it stays closed on app
launch (the dock needs to exist in the Widget hierarchy for its state
to be restored).
Alas, outside of Windows, InitBrowserPanelSafeBlock uses a separate
local QEventLoop to block the main thread while still allowing UI
events to be processed to launch a CEF instance.
By this point in the code execution, the primary event loop has not
been started yet, so the event loop launched by
InitBrowserPanelSafeBlock temporarily becomes the main application
event loop, which initializes all Widgets, finds no active window
state for the widgets, and thus treats them as "visible", calls the
showEvent method on every browser dock, which thus loads the associated
websites.
The dock state is restored after all that, which leads to each browser
dock being "shown" (even though the main QApplication hasn't even
started yet), the associated sites are running (including audio and
video output) but then hidden again, which leads to surprising audio
output seemingly coming from "nowhere".
All browser docks call the browser initialization methods synchronously,
which has the benefit of not spinning up a premature event loop, and does
not trigger Qt view state changes before all Widgets have been
initialized. Having the YouTubeAppDock behave the same does thus not
negatively impact UX.
Some errors include HTML links directing users to e.g. driver updates
or further information. Using a raw newline instead of <br> causes Qt to
skip parsing the HTML, resulting in an ugly mess of HTML displayed to
the user instead of the intended links.
The native popup has the problem that it doesn't work for links, leaving
us with an unclickable text where a link should be. Qt 6.6 has an option
to disable the native dialog, so let's add this to make the link
clickable again.
Co-authored-by: derrod <dennis@obsproject.com>
Moving the cleanup to OBS_FRONTEND_EVENT_EXIT in #8735 only handled the
cleanup from the dockable window, as the regular stats window is deleted
on close when the UI is shut down. This caused an event handler leak
each time the window is manually closed, resulting in crashes. This code
looks a bit wrong since we delete the same handler in multiple places,
but this is due to the code being used by both the dock (non-closable)
and the window (closable). The OBS_FRONTEND_EVENT_EXIT handler handles
cleanup from the dockable stats window, and the window close handler now
handles cleanup from the non-dockable stats window.
ExpandCheckBox was introduced in 88b6c63, but the seemingly replaced by
SourceTreeSubItemCheckBox during development. This means that it became
completely unused.
Like many IDEs, Xcode has this feature where it shows the declaration of
the method currently being worked in. However it gets confused by scopes
starting inside of preprocessor guards and ending outside them,
resulting in the declaration of OBSBasic::ReceivedIntroJson always being
shown in window-basic-main.cpp from that method downwards. We can work
around that by starting and ending the if-scope outside of the ifdefs.
The current regex pattern will successfully match if the cache variable
contains the pattern anywhere inside the string. Let's restrict the
regex pattern such that it requires the pattern to be precisely between
the beginning and end of the string with no other characters in between
those anchor points.
The OBS client side timeout for the YouTube API is too short. New
changes proposed to the YouTube API will cause this to fail for
creating a broadcast. Increase to a minimum of 60.
In the settings window, when the user chooses to use a stream key
instead of connecting their account, it switches widgets to show the
stream key UI. However, when opening the settings window again after
having done that, it's supposed to continue to show the stream key UI,
and that functionality was broken by obsproject/obs-studio#9272. It did
this because OBSBasicSettings::ServiceChanged() no longer called
reset_service_ui_fields() because the lastService member variable was no
longer set to empty in OBSBasicSettings::LoadStream1Settings().
This fixes it by just adding a parameter to
OBSBasicSettings::ServiceChanged() to make it forcibly reset the fields
when loading stream settings.
Some hardware encoders fail in indecipherable ways if you try to at
certain low resolutions such as 17x17 or 19x19, but works at 20x20.
However, there is no good way to determine what the minimum working
resolution is for every possible encoder. There isn't that much
difference between 8x8 and 32x32, except that the latter will reduce or
eliminate such odd failure cases, so let's increase the minimum required
output resolution by just a little bit.
YoutubeAuth::chat is a raw pointer that is uninitialized. Most of the
time it doesn't matter, since the object it points to is created at the
application start. However, in case of Wayland (Linux), CEF is not
initialized (because it's not supported) and the chat object will never
get created.
This patch fixes crash on Wayland by initializing `chat` to `nullptr`.
Co-authored-by: Roman Podoliaka <roman.podoliaka@gmail.com>
This reverts commit aeed4a3aa1.
The source type `browser_source` is now available for all supported
platforms so we don't need to fallback to `linuxbrowser-source`.
The status bar message was not vertically aligned properly to
other widgets in the status bar. We have to implement our own
message system here, as the default status bar in Qt has hardcoded
paddings.
CEF apparently doesn't like it and is unable to handle when you
destroy/recreate CEF instances very quickly, so... let's just put a
timer on this insanely terrible function. I guess. Whatever.
If you're using an encoder from a plugin, and you remove that plugin,
this value will be null, causing a crash. (Lain note: I realize that
this isn't the best solution to this problem, though there really isn't
any way to infer what type of encoder the prior encoder was. Should also
save the encoder type as additional setting metadata in future
versions.)
After commit 8955338624, 'UI/update/models/whatsnew.hpp' is included
unconditionally, even with 'ENABLE_WHATSNEW=OFF', which breaks
building OBS Studio without nlohmann JSON.
The use of GetCurrentSceneItem can lead to a race condition between
the graphics thread (where this function will be run) and the main UI
thread, as both access the scene list model and iterate through its
members which can change during iteration (as there is no lock around
its access).
Firstly, removes the UpdateTransformShortcuts() method introduced in
c33fa8b which was trying to re-implement the behavior of
UpdateEditMenu().
Secondly, updates UpdateEditMenu() to account for sources without video.
Sources without video shouldn't be able to have their transform edited,
copied, pasted, or changed in any other way (because they don't have
one).
Attempting to set the window title of a BrowserDock that is closed, then
opening the BrowserDock, would show that BrowserDock with an incorrect
title. We can handle this by overriding the showEvent of BrowserDock and
manually setting the window title after the showEvent is called,
hopefully ensuring that we are only setting the window title on a window
that exists.
Instead of relying on the dock's having their window title already set
for the menu items to be updated, let's just manually update the menu
item text ourselves when updating the item.
This is a workaround for a change in Qt behavior between 6.5.1 and
6.5.2[1] which affected the outcome of attempting to set the window
title of a window that is not currently open. In Qt 6.4.3 and 6.5.1, the
window title would be set. In Qt 6.5.2, the window title will become
NULL/empty.
Instead of relying on Qt to have valid titles for custom browser dock
windows that we manage, let's track the names ourselves.
[1]: c153066baa
Works around a bug in Qt where setWindowTitle called on a native window
that hasn't been initialised yet causes the title data to be lost,
resulting in the window's title never being set.
The resize event was triggered during the shtudown of OBS. A commit
9f330050ef moves the actual resize code in `gl_update` to the main
thread without synchronizing between the main thread and the graphics
thread. When the display context is destroyed just after the resizing,
the delayed resize code can read the destroyed display context.
As a workaround, this commit will destroy the display context at the
earlier timing of the shutdown process.
The display context was sometimes created again by a resize event after
destruction from `OBSBasic::closeEvent`. That resulted in the display
context alive until reaching the destructor of the `OBSQTDisplay` class
and caused a crash in `render_display` on macOS. To avoid this, destroy
the display context at the event `SurfaceAboutToBeDestroyed` and let not
create the display context again.
The commit e67fdbca79 is a workaround for QtWayland bug that the
`QEvent::Expose` was never received on a certain condition.
The bugfix was introduced to QtWayland 6.1.0 so that the workaround is
not necessary for now.
Like in 4f8b17d61, "toggled" is the signal that listens for changes
however they may have happened (including through things like
accessibility software), whereas clicked only is emitted on mouse
clicks. We should react to any kind of change, whatever prompted it.
Capping the precision for whole numbers results in scientific notation
being used for any number that's larger than what can be displayed with
the given precision. In case of 2, that would be anything larger than
99, resulting in 120 for example to be shown as 1.2e+02.
Let's not cap the precision for whole numbers ourselves and just use the
default value (which at the time of writing would be 6).
The QGroupBox::clicked is specifically for mouse clicks. This means that
it's not emitted if the "checked" property is modified through other
means than the mouse, for example programatically or through
accessibility software like VoiceOver.
However, we do want to catch such events, so the QGroupBox::toggle
signal (which as per the Qt documentation is the notified signal for the
"checked" property) is the appropriate one to use.
When loading a settings pane, "loading" should be set to true to stop
signals being emitted by settings changes to trigger while the settings
are loading. This should have always been set here as well, but wasn't
detected so far because currently there are no settings in accessibility
that have signals connected that track changes.
Commit 85f9a8661b did not take into account whether YouTube was using a
stream key or not, it would prevent the auto-configuration dialog from
being able to save a YouTube stream key if the user opted to enter in a
stream key instead of connecting their YouTube account.
This simply fixes that by checking to see whether or not the user fully
connected their account or not.
Fixesobsproject/obs-studio#6406
Macros can be hard to read and are usually not very friendly to use in a
debugger. Using a static function instead gives the advantage of better
syntax highlighting in IDEs and better debugger support.
Due to a particular... plugin we all know and love, the new
QCefWidget::executeJavascript() function may not actually be present as
it should be. So, we have to verify that it's available, and only use it
in that case.
Apparently the YouTube app dock thing didn't take into consideration the
fact that the user may have the browser source available, but may not
have browser docks available (via the `cef` global). This is because
certain system configurations do not support browser widgets, such as
Wayland on Linux.
(Lain note: There's multiple things I'm unhappy about here. I'm mostly
going to blame myself for multiple reasons.)
This reverts commit 58fb63030c.
This is not the ideal fix for the fact that scripts can use a deferred
update flag because it makes it so that filters in general cannot use
flags. We need to disable this via other means.
The alias is not really helpful and sometimes confusing, e.g. for HEVC
(without libx265) it ends up being "hevc_nvenc (hevc_amf)" since it just
contains the name of the default encoder. So instead of using the name
of the default encoder, show the full name of the encoder instead.
Previously we left the QSV -> x264 in the fallback function to deal with
the autoconfig wizard possibly selecting QSV outside of the UI. But this
actually runs for both recording and streaming every time obs starts,
reseting user's config if they selected QSV for recording.
Instead avoid recording from being selected in the UI and by the
auto-config wizard, and revert changes to the fallback function.
Same as previous commits, flipped sceneitems in bounding boxes ought
to be treated as not flipped for the purpose of UI interaction in
order to function correctly.
Flip state of the item should be ignored when drawing guides for items
in bounding boxes, as the UI uses the countour of the bounding box, not
of the flipped item in it.
The preview interaction and decoration of items inside bounding boxes
is based on the bounding box state, not on the item's state. As such,
the flip state of the item should be ignored when drawing and finding
the rotation handle, since the bounding box itself isn't flipped, only
the item inside the bounding box.
The angle at which the rotation handle is drawn is independant from the
sceneitem's flip. As such, this changes `FindHandleAtPos` to use the rotation
angle of the item, free of its flip, instead of doing post-rotation
corrections. This fixes a bug where when the item was horizontally flipped,
the handle was not "found" at its drawn location.
Fixes a bug where when you clicked "Show more..." in the service
settings combobox, the currently selected service would be changed to
the first service in the list.
A small refactor of AddProjectorMenuMonitors, as template functions need
to be fully defined in header files.
Also makes the function static as there is no point of it being a
non-static member function. In theory it could also be moved out of
OBSBasic but there isn't really a better place for it so let's leave it
where it is for simplicity.
This changes the return type for getting the filter index from a
size_t to an int. This makes it easier for developers to use,
as an invalid index just returns a -1.
This redesigns the status bar to look more modern.
Co-authored-by: Warchamp7 <warchamp2003@hotmail.com>
Co-authored-by: Georges Basile Stavracas Neto <georges.stavracas@gmail.com>
The frontend API was never explicitly cleaned up which has lead to
several crashes from plugins continuing to call it even after OBS itself
has been destroyed. We now destroy the API after the exit event, so
further calls by plugins will be no-ops instead of potentially accessing
destroyed OBS state.
This also required some changes to our own use of the API, as we relied
on the OBSBasic destructor to clean up some callbacks, by which point
the API should have already been destroyed.
Linux QSV drivers have varying capabilities depending on if the
free/non-free Intel drivers are installed. This means we cannot safely
expose QSV in simple mode as users may be unable stream without changing
to advanced mode.
The macros `YOUTUBE_ENABLED`, `RESTREAM_ENABLED`, and `TWITCH_ENABLED`
are not defined when the associated features are not enabled.
When the code was originally implemented, there were definitions to set
`0` if the feature is disabled. However a later commit 349372b removes
the definition. Now, the use of `#ifdef` is better than just `#if`.
Instead of actually removing the file, simply rename it to the .bak
version which OBS ignores in most cases. That allows users to recover
accidentally deleted collections more easily, while only taking up a
few kilobytes of disk space at most.
A bug which was technically introduced in df446c3f6e: the author
assigned video mixes to the virtual camera and did not realize you have
to wait for full capture stop with the "deactivate" signal rather than
the "stop" signal. This situation is understandable because these
signals are likely confusing and need more documentation.
The way an output works when it stops is it has three stages:
- "stopping", which is the moment the user themselves stop the output
- "stop", when the output's implementation has stopped and returns
success or an error code
- "deactivate", when the output inself (not the implementation) has
fully stopped capturing encoded or raw frame data after the
implementation has stopped. This is done in a separate thread,
end_data_capture_thread(), for performance and data race reasons, and
is when it's "actually fully stopped for real this time"
(Lain note: I sincerely apologize for this confusing signal design)
The author of df446c3f6e was likely confused as to why they could not
destroy the video mix in the "stop" signal without triggering a race
condition in end_data_capture_thread(), and instead decided to make a
workaround to clear the video mix set to the output before destroying
the video mix. Unsetting the video mix I'm guessing *seemed* to fix the
problem for them at the time, and destroying the video mix separately
after that automatically stops capture, so it *technically* worked
because the deactivate thread cannot be called until the
implementation's stop signal has executed. However, it was still the
incorrect way to handle the problem, because it circumvents the output's
mechanism for deactivating the frame data capture by destroying the view
instead.
The reason this was the incorrect way to handle the problem became
exposed in 7cd7ca80f8, when tytan made it so it uses the main video mix
under certain circumstances. The main view is special and cannot be
destroyed, so the mechanism to stop frame data capture that the
virtualcam was using before failed, and raw frame data capture continued
perpetually until the end of the program erroneously.
This fix solves the bug by using the "deactivate" signal rather than the
"stop" signal, thus allowing the output to fully end its data capture
and *then* destroy the video mix.
Fixesobsproject/obs-studio#9153
Unobfuscated secrets require the hash values to be set to 0, by default
CMake will treat 0 as a falsy value. This commit adds support for
_either_ 0 _or_ a valid hexadecimal hash.
YouTube Live now supports 5.1 surround sound.
The surround sound warning in Settings has been updated to reflect that.
Signed-off-by: pkv <pkv@obsproject.com>
When changing audio channels, sample rate, or audio buffering settings
multiple times by hitting `Apply` button, the internal `changed` state
got corrupted. To avoid this, set the changed state when changed
something, but do not clear the changed state.
Creation of Info.plist files through Xcode is more canonical and
future-proof, as it will automatically pick up changes/updates
introduced by Apple. Non-standard keys can still be added via a
template file, which will then be extended by Xcode with the default
keys.
The branch is never taken and the code acts as documentation for a
(possible) future use case. According to YAGNI it could be removed
entirely, wrapping it in a comment is the less destructive approach.
This provides the UI glue to enable the WHIP service introduced in the
obs-webrtc plugin.
Co-authored-by: John Bradley <jocbrad@twitch.tv>
Signed-off-by: pkv <pkv@obsproject.com>
The configuration directory was $HOME/.obs-studio until version 0.10.1
and then moved to XDG's configuration path by ba02e065fe.
The function move_to_xdg was introduced to move the old configuration
directory and left almost untouched for 8 years.
Some plugin does that by deleting the QAction returned by
obs_frontend_add_dock().
Now that obs_frontend_add_dock() is deprecated,
obs_frontend_add_custom_qdock() replace this usage.
before: widgets docked on the sides of the preview are of the height of
the preview
after: the user can toggle the setting in the docks menu so the side
widgets are of the height of the whole window
UI: change menu label to Full-height docks
This fixes the shortcuts not showing up in the filters context
menu for rename and remove. This also has a code cleanup for
the rename action, as now it is created in the ui file.
This change allows the virtual camera to really output what is in the
program view, some plugin interract with this view but their changes
does not appear on the virtual camera.
Switching to a static library that contains version information as
const char strings has multiple benefits:
* The version information provided externally via compiler definitions
will fail compilation early if malformed
* An updated version string (which will happen with every commit) will
not invalidate existing compilation units, because only the static
library is affected by the change
* An update of the version change just requires a recompilation of the
static library and a linker update
* An update of the version will _not_ infect the rest of the codebase
(as it does currently, because everything includes obsconfig.h one
way or another)
* Other modules which used the macro definition directly have been
updated as much as possible to use the proper getter method from
`libobs` instead (some Windows-specific modules use preprocessor
string composition, the value has been added as a compiler definition
directly in those cases)
* Because the impact of a version change due to a commit hash change
is limited to the static library, ccache hit rates should be
improved considerably
Deferring properties was only ever made for input and transition
sources. As other property dialogs do not have an "Ok" button that would
cause an update to happen, if the deferred flag was set the callback
would never be called.
Also clarifies the docs to reflect this.
Fix a bug where if a source's width or height becomes 0 while you're
stretching the size of a source or modifying it from the transform
dialog, it would become permanently invisible
Closesobsproject/obs-studio#7962
