Users on Wayland are displeased that they cannot see their hotkey
bindings. This enables key reporting like X11, and has the infrastructure
in place in case Wayland ever decides to allow for capturing input.
Add an extension point for org.freedesktop.LinuxAudio.Plugins,
and create the corresponding folder at post-install time.
Commit originally made by @hfiguiere for Flathub.
This is what the Flathub manifest does, which allows for plugins
distributed through Flatpak extension points.
Add extension point for Flatpak-distributed plugins.
In the future, a full FFMPEG build with x264 and nvenc will be
provided as an extension to the org.freedesktop.Platform runtime,
but for now we have to manually build it.
Update x264 to the latest git commit.
* aja: Initial commit of AJA capture/output plugin
* aja: Fix clang-format on aja-output-ui code
* aja: Remove script used during dev/testing
* aja: Address pull request feedback from @RytoEX
* aja: Remove the SDK sources and update CMakeLists to point to new headers-only/static libs dependency distribution.
* aja: Only build AJA plugin on x64 on macOS for now
* aja: Remove the non-English placeholder locale files. The english strings/files will be produced via crowdin, according to @ddrboxman.
* aja: Add FindLibAJANTV2.cmake script to locate the ajantv2 headers and static libs in the OBS external deps package(s). Tested on Windows x64. macOS and Linux x64 TBD.
* aja: Add ajantv2/includes to FindLibAJANTV2 include search paths
* aja: Remove commented code from aja CMakeLists
* aja: Remove debug code and comments that are no longer needed.
* aja: Fix indentation
* aja: Remove disablement of clang-format in routing table and SDIWireFormat map
* aja: Use spaces for all indentation in widget crosspoint arrays where we disable clang-format
* aja: Address code style comments made by @RytoEX
* aja: Fix uneven indentation
* aja: More fixes to if/else placement and remove superfluous comments.
* aja: Rename 'dwns' to 'deactivateWhileNotShowing' for clarity. The DeckLink plugin still uses the variable name 'dwns' and should be changed, if desired, in a separate PR.
* aja: Remove X11Extras dependency from AJA Output frontend plugin
* aja: Add patch from Jim to find AJA release/debug libs
* aja: Improve AV sync of queued video/audio sent to the AJA card in the AJA Output plugin.
The flatpak-builder tool now supports passing secrets options to the
build system. These options are not printed during the build, nor
added to the resolved manifest after build, so they don't leak env
vars from CI.
Make secret variables part of the Flatpak workflow environment, like
the main workflow. Pass the various services hashes and clientids to
the build system using the new "secret-opts" key.
(Jim note: Rather than copy the QtNetwork library manually like we were
doing before, this makes it so that QtNetwork is used as a dependency of
the UI. The cmake used to copy the library manually thus us no longer
necessary.)
This reverts commit c370692773.
The PR that introduced this commit was opened before the browser plugin
was enabled, and when the browser was effectively enabled, this change
was already in place and didn't conflict, leaving us with both a
-DBUILD_BROWSER=ON *and* -DBUILD_BROWSER=OFF in place.
Fixes an issue where on macOS the app’s Info window in Finder would say
"Version: OBS - Free and Open Source Streaming/Recording Software"
instead of the version set in CFBundleShortVersionString.
Since CFBundleGetInfoString is deprecated anyways, it can safely be
removed.
It is not possible to run host system executables like modinfo, pkexec,
and modprobe inside a Flatpak sandbox. However, Flatpak provides a way
to run command on the host system: the flatpak-spawn executable.
flatpak-spawn is a tiny helper that, when executed with the '--host'
parameter, talks to the org.freedesktop.Flatpak D-Bus interface to run
and retrieve the return value of the executable. This provides OBS Studio
a way to escape this sandbox limitation without opening large holes in
the sandbox.
Make v4l2's implementation of VirtualCam run system commands using
flatpak-spawn when inside a Flatpak sandbox. The detection of the sandbox
is done by checking the existence of the /.flatpak-info file, which is
created by Flatpak itself, and only exists inside the sandbox. If OBS
Studio is not running inside a Flatpak sandbox, run the exact same command
it used to run before this commit.
Add the permission to talk to the org.freedesktop.Flatpak D-Bus interface
to the Flatpak manifest, so we can run flatpak-spawn with the '--host'
parameter.
Notice that the same constraints apply with and without Flatpak: the host
system needs to have the v4l2loopback kernel module available for the v4l2
implementation of VirtualCam to work.
It seems that the browser source works inside a Flatpak sandbox and/or
Wayland with CEF 4280, so let's try and reenable it.
This reverts commit e64c61710f.
Unfortunately, neither Ubuntu 20.04 nor 18.04 have a recent enough
PipeWire package. Disable the PipeWire bits of linux-capture there.
The Flatpak workflow is still able to build it, so keep it enabled
there.
luajit developers ask people to use branches instead of tarballs, however,
Flatpak interprets having both 'commit' and 'branch' fields as 'use this
branch, and this commit should be at the top of the branch', which is not
really what we want.
Remove the specified commit from the Flatpak manifest.
- Update LuaJIT according to upstream recommendations (maintainers of
this project prefer that distributors use the git repository directly,
instead of archives. Do that, and also update the Lua path for swig.)
- Update ffmpeg and drop patch which is included in this updated release
now.
- Update x264
- Update nv-codec-headers
- Update mbedtls
The Chromium sandbox is conflicting with the Flatpak sandbox in a non-trivial,
non-workaroundable way. Until Chromium / CEF provides a way to unconditionally
disable the entire sandbox, let's not degrade the Flatpak experience.
Co-authored-by: lvsti <lvsti@users.noreply.github.com>
Co-authored-by: Sebastian Beckmann <beckmann.sebastian@outlook.de>
Co-authored-by: Stefan Huber <sh@signalwerk.ch>
Co-authored-by: Ryohei Ikegami <iofg2100@gmail.com>
Co-authored-by: Colin Dean <colin.dean@target.com>
Co-authored-by: Wolfgang Ladermann <extern.ladermann_wolfgang@allianz.de>
Co-authored-by: Simon Eves <simon.eves@omnisci.com>
Co-authored-by: Colin Nelson <colnnelson@google.com>
Co-authored-by: Yoshimasa Niwa <niw@niw.at>
Co-authored-by: Michael Karliner <mike@modern-industry.com>
Co-authored-by: Jason Grout <jgrout6@bloomberg.net>
Co-authored-by: Alfredo Inostroza <jadenguy@gmail.com>
Co-authored-by: Daniel Kennett <daniel@cascable.se>
Co-authored-by: Gary Ewan Park <gep13@gep13.co.uk>
Co-authored-by: José Carlos Cieni Júnior <cienijr@outlook.com>
'PlugIns' is normal for MacOS and whatever is generating qt.conf
(macdeployqt?) is using 'PlugIns', so this both builds and generates
an app that works on case insensitive filesystems but fails to load
on case sensitive filesystems.
I haven't tested building the installer, only modifying the
qt.conf in the installed copy of OBS.app.
Fixes https://github.com/obsproject/obs-studio/issues/2302
The current scripting directory path given to Python on OSX is
`../data/obs-scripting`. This works when bin/ and data/ are in the same
folder like rundir after a compile and in the old .app bundles but the
new normal .app bundle structure is different.
For bundles move obspython.py and _obspython.so to the binary folder,
just like the Lua one, and pass that bin path to the Python interpreter
so it can find it.
The current working dir isn't guaranteed to be set with .app bundles
so look up binary path at runtime.
As noted by ePirat, the VLC releases on GitHub are not official
releases, just repo snapshots. Use the official VLC 3.0.4 release
tarball to ensure maximum stability.
AppVeyor recently updated their Qt version to 5.11.2 which breaks
builds, because of the outdated path. Using this path should always
ensure that the newest subversion of Qt 5.11 gets used.
On Ubuntu 16.04 (Xenial), we have to build mbedTLS from source because
the version available from apt, mbedTLS 2.2.1, does not work with the
RTMPS implementation in OBS, which requires mbedTLS 2.7 at minimum. This
is due to the use of the function mbedtls_ssl_conf_dh_param_bin, which
was introduced in mbedTLS 2.7.
Build on Linux (Ubuntu 16.04) with AppVeyor. Disable PowerShell on
Linux. Ubuntu 16.04's dependencies are different from the existing
Ubuntu 14.04 dependencies, so a separate script for installing those
dependencies has been added.
Move the AppVeyor install scripts to a separate script file. This
shortens the appveyor.yml file and makes it easier to introduce new
install scripts for other build environments.
There is one minor change to the install script, and that was to change
directory back to the obs-studio git root directory.
(This also modifies the CI scripts accordingly)
It was difficult to build local obs-browser builds on macOS due to the
way the frameworks work and libraries are linked. This solves that
issue so that local builds of obs-browser should always work fine.
Pull Request #1420 changed the URL for the macOS deps package. When it
did that, it also changed the URL from HTTPS to HTTP. Since this asset
is available over HTTPS, we should use that.
* Add mbedTLS in brew install
* Update Sparkle from 1.16.0 to 1.20.0
* Update Opus from 1.1.3 to 1.2.1
* Update libogg from 1.3.2 to 1.3.3
* Update libvorbis from 1.3.5 to 1.3.6
* Update libvpx from 1.6.0 to 1.7.0
* Update Jansson from 2.9 to 2.11
* Update FFmpeg from 3.2.2 to 4.0.2
This fixes the pkg build of obs-browser by properly copying CEF to
`OBS.app/Content/Frameworks/Chromium Embedded Framework.framework`,
rather than `OBS.app/Content/Frameworks`.
The log output of the travis CI for macOS is unreasonably large. This
reduces the verbosity of the output and reduces the size of the build
logs significantly.
