With v5, NDI doesn't compile because of NDIlib_frame_type_source_change
missing. According to the NDI license, the most recent version should
be used, anyways.
instead of building own Spout, use pre-built release on GH
This actually reverts the commit b98b5ab0 from 2023-11-06.
Date: Mon Nov 6 09:08:46 2023 +0100
Make the bug <https://github.com/CESNET/UltraGrid/issues/478> fail
with distcheck.
The arm64 Homebrew uses different location than x86_64 so that it didn't
catch the problems (Homebrew libs were still visible).
Co not continue if SDK_NONFREE_PATH is set but deltacast
SDK not present there. At this time, this will fail later in
.github/scripts/install-common-deps.sh, anyways. Also the DELTACAST
SDK remains the last non-free SDK so it doesn't make sense to define
SDK_NONFREE_PATH GH variable while not having that SDK.
compute SHA256 checksums for builds
this is useful at least for 2 reasons:
1. user can ensure that the build comes really from the CI (isn't
tampered)
2. can identify or match version with appropriate action run (the $VERSION
is usually always "continuous", except for releases)
The sha256 sum is currently only show in the protocol, not uploaded
to GH releases - this doesn't seem to be needed because GH currently
already prints the checksum by itself on the download (releases) page.
Skip computing checksum in Win when uplad-artifact is used - the zip
is actually created by the Action so we cannot compute the checksum
on it.
/sbin/sha256sum is not present in CI mac workers for some reason
Remove commands that actually do nothing because uv-qt.exe is installed
by make install and the contents of bin/ subdirectory is not copied
"as is" as the command seem to imply.
+ use $(TARGET) in Makefile instead of bin/uv (which will be bin/uv.exe
in MSW but obviously it worked even without the extension)
- coverity-scan.yml: typo
- configure.ac: misleading message if SDL not found -> do not mention
version at all, it was perhaps meant "didn't found any version" but it
might have also been that did't found specified version...
- configure.ac: removed sdl_mixer from summary
This reverts 960d2315 (2022-02-21) because the commit 18f049608
(2022-04-20) already restricts the scheduled run to CESNET Git repository
or manual push.
So that the token unavailability should be considered error and thus the
condition doesn't need to be re-iterated in every step.
in similar fashion as already done for install_sdl.sh
As a bonus, we do not need to depend on whole prepare.sh for cache
rebuild but only on this script, which is perhaps cleaner and may
yield less rebuilds. Also prepare.sh is slightly easier.
add all to one script with deps/build_install/install_cached actions
- yml - use action @main
- cache: do not depend .github/scripts/Linux/prepare.sh, just the
catch-all SDL install file
- build SDL,SDL_ttf without entering the dir (as fluidsynth already does)
Print the output variable assigned in the script in case that this
fails, otherwise it on the output without NL and will be prepended to
next var (if any).
This also simplifies the call in YAML.
see the previous commit
This changes the manual handling introduced by the commit 925db63e. Also
$FEATURES are not passed explicitly as added in the commit 9eb29225.
Instead of use "Cache XY" use more techical "Run action/cache for XY"
(similarly as implicit name when "name" key was not used).
Although the action is really actions/cache, the problem with the previous
name is that it actually doesn't cache the named item. It checks, and
if found loads, the item. But even if the item is not found, the action
reports success in GH web UI, which is quite misleading. So using rather
neutral name for the step should slightly decrease the confusion.
Use runner ubuntu-22.04 instead of ubuntu-latest. Because ubuntu-latest
has switched to ubuntu-24.04 recently, whiich breaks the build for arm64.
There seem to be an unresolved bug in QEMU >= 8.1
(U24.04 has v8.2.2) when using ARM64 and bullseye:
<https://gitlab.com/qemu-project/qemu/-/issues/2377>.
(ARM64 buster and bookworm, as well as ARMHF, may be fine as indicated
in the bug description.)
This can be reverted later (either when fixed in QEMU or when switched
the chroot to bookworm). But it shouldn't matter much, anyways, since this
is just the "host" environment.
Also include runner ver in chroot key because it seem to make a difference
from the above..
the package `qemu` is no longer present in U24.04, which is now
ubuntu-latest default Ubuntu version as a GH runner. But we don't need
it, we use just the static version (qemu-user-static).
suppress `brew list` output
It is used just to check if a package is installed and the output is
print is output to stdout. To stderr is print that it if the package is
not installed.
The file name has changed and --enable-deltacast was used just if the
download succeeded (which didn't).
Fixes the commit 98d4969b (2024-05-24) that disabled deltacast, because
the archive was not downloaded due to wrong name.
macOS 12 is no longer supported and as Homebrew has a policy, that it
compiles packages from sources on unsupported macOSes, it is needed to
switch to newer macOS to keep reasonable build times.
the message: `Error: No such keg: /opt/homebrew/Cellar/asciidoctor`
is now print on stderr, which disturbs GitHub CI (complains about it in
action summary annotations; just a warning).
`
Some features that are used by x86_64 Linux build are disabled, for the
rest dependencies are installed.
disabled:
- CUDA features (except of NV Jetson usually there is no CUDA)
- lavc vdpau (similar to CUDA, it is primarliy NV-focused API)
- Qt GUI
- Cineform (doesn't compile on ARM)
dependencies installed:
- libsdl2-ttf-dev
- libsdl2-mixer-dev
- libva-dev
Since we must now build the new libajantv2, which takes some 3 minutes,
cache the build in order not to increase the build time.
Also use libajantv2 path for the SDK instead of AJA (doesn't require
unneeded rename), which will be Windows specific.
Only do it in MSW, since the build is slowest. But it can be easily used
also for other platforms.
\+ error from the notarize script if called
Previously (since commit 1c7c3ac1), the script silently exitted without
cigning and notarizing the app, which may harden debugging. Now the skip
is explicit and if the script is called anyways, it returns error.
enhances the commit 1c7c3ac1 (2024-03-22)
