I did some tests on my own fork, and we're writing to ~/.transifexrc but
the client can't seem to read that file. maybe issue with $HOME or
something.
Workaround is to set TX_TOKEN environment variable and now the pesky
~/.transifexrc file is not needed.
Currently we're using the python client which uses an API that they
state will sunset Nov 30, 2022.
`tx push -s` actually appears to work properly, some of the other
commands require tweaking, like instead of suggesting `tx pull -a` in
dist/languages we need to suggest `tx pull -t -a`
Using MinGW in the future may not be ideal as it does not work very well
with crash dumps (#8682).
Switch back to GCC on MinGW. This also gives CI a way to check GCC 12
(as of writing, or whatever version of mingw-gcc Arch happens to be
shipping on a given week).
Recent versions of Docker appear to cause the Qt linuxdeploy plugin to
throw a boost file copy error.
This switches from linuxdeploy to a script of mine I've been working on
for a while.
The current AppRun is more difficult to update. This script still
uses the old version of AppImageKit-checkrt, but now we use the shell
script version so we can set our own environment variables as the
application starts up.
This specific version searches for and sets the correct root CA file to
prevent SSL errors in yuzu.
Instead of including yuzu and all the sources it uses directly, include
only what specifically belongs to yuzu. Submodules can be downloaded
separately later using git since a shallow clone includes minimally all
the repository information needed for it.
The Early Access AppImage needs to be accessible through liftinstall, so
a couple modifications need to made:
The DIR_NAME needs to not include the revision info.
The EA AppImage name cannot contain revision info.
The EA AppImage has to be packaged with the rest of the yuzu package,
which means both binaries and the source are bundled with it now in an
archive.
In addition, fix the source archive so yuzu can actually be built from
it.
upload: Copy AppImage to both mainline and EA release package
vcpkg: Add Catch2 2.13.9
Catch2 >= 3.0 is not compatible with earlier versions, and for now we
must override the desired version in our vcpkg manifest. We can do this
programmatically by using VCPKG_MANIFEST_FEATURES.
CMakeLists: Search for lz4 CONFIG mode first
vcpkg's lz4 CONFIG cmake script works in Release mode but not in Debug
mode, failing to copy the correct DLLs at compile time.
We still need to search for the regular mode for system-installed
versions.
CMakeLists: Clean up boost exports
Remove some Conan-specific workarounds.
CMakeLists: Use vcpkg for MSVC by default
Not enabling it generally since it's much easier to have system
dependencies installed for Linux and MinGW.
[REUSE] is a specification that aims at making file copyright
information consistent, so that it can be both human and machine
readable. It basically requires that all files have a header containing
copyright and licensing information. When this isn't possible, like
when dealing with binary assets, generated files or embedded third-party
dependencies, it is permitted to insert copyright information in the
`.reuse/dep5` file.
Oh, and it also requires that all the licenses used in the project are
present in the `LICENSES` folder, that's why the diff is so huge.
This can be done automatically with `reuse download --all`.
The `reuse` tool also contains a handy subcommand that analyzes the
project and tells whether or not the project is (still) compliant,
`reuse lint`.
Following REUSE has a few advantages over the current approach:
- Copyright information is easy to access for users / downstream
- Files like `dist/license.md` do not need to exist anymore, as
`.reuse/dep5` is used instead
- `reuse lint` makes it easy to ensure that copyright information of
files like binary assets / images is always accurate and up to date
To add copyright information of files that didn't have it I looked up
who committed what and when, for each file. As yuzu contributors do not
have to sign a CLA or similar I couldn't assume that copyright ownership
was of the "yuzu Emulator Project", so I used the name and/or email of
the commit author instead.
[REUSE]: https://reuse.software
Follow-up to 01cf05bc75
With our recent switchover from conan to vcpkg, we're shipping a few
more dll files, these need to be in the full zip.
cp .\build\bin\*.dll .\artifacts\
also tacking on the fix where we're shipping scm_rev.cpp accidentally
The slim docker container that runs transifex needs a few packages added
in, curl zip unzip
I've tested everything except actually pushing to transifex, but it's
not November 2022 yet so we're fine for now. Or we're actually using the
newer client and all is well.
This is related to 8486
Ninja places the exe files into .\build\bin while MSBuild may place them
into .\build\bin\Release
upload.ps1 was originally written for use with Azure Dev Ops to cough up
about 5 files and the script appears to be used for both CI and
mainline builds
GHA (GitHub Actions) makes available a single zip of the items uploaded by
each Upload action (artifacts directory), so we want to work with that.
I'm doing changes to upload.ps1 to accomplish this.
The changes to the verify.yml are as follows
-DGIT_BRANCH=pr-verify changes the header in yuzu, instead of saying
HEAD-<hash>-dirty it'll say pr-verify-<hash>
-DCLANG_FORMAT_SUFFIX=discordplzdontclang tricks the CMake stuff for
discord-rpc to NOT run clang-format, as this was marking CI builds as
dirty
I'm also making it upload just the exe by itself, as the msvc builds are
quite chunky. but maybe this is unnecessary.
Currently the MSVC artifact option is a 274MB zip that contains 3 copies
of the DLLs, and 4 copies of the source tarball, and zero copies of yuzu.exe
This PR should have msvc artifacts of about 190MB that downloads as 81 MB zip
Variables in question:
AZURECIREPO TITLEBARFORMATIDLE TITLEBARFORMATRUNNING DISPLAYVERSION
CMakeModules/GenerateSCMRev.cmake has some logic that looks at BUILD_REPOSITORY variable inside CMake
src/common/CMakeLists.txt has some logic that takes some items from environment variables and
sets variables inside CMake
This is the whole section at the moment.
if (DEFINED ENV{AZURECIREPO})
set(BUILD_REPOSITORY $ENV{AZURECIREPO})
endif()
if (DEFINED ENV{TITLEBARFORMATIDLE})
set(TITLE_BAR_FORMAT_IDLE $ENV{TITLEBARFORMATIDLE})
endif ()
if (DEFINED ENV{TITLEBARFORMATRUNNING})
set(TITLE_BAR_FORMAT_RUNNING $ENV{TITLEBARFORMATRUNNING})
endif ()
if (DEFINED ENV{DISPLAYVERSION})
set(DISPLAY_VERSION $ENV{DISPLAYVERSION})
endif ()
Between packages breaking, Conan always being a moving target for
minimum required CMake support, and now their moves to Conan 2.0 causing
existing packages to break, I suppose this was a long time coming. vcpkg
isn't without its drawbacks, but at the moment it seems easier on the
project to use for external packages.
Mostly removes the logic for Conan from the root CMakeLists file,
leaving basic find_package()'s in its place. Sets only the
find_package()'s that require CONFIG mode as necessary. clang and linux
CI now use the vcpkg toolchain file configured in the Docker container
when possible.
mingw CI turns off YUZU_TESTS because there's no way on the container to
run Windows executables on a Linux host anyway, and it's not easy to get
Catch2 there.
Uses the MinGWClangCross toolchain script to build yuzu. Disables our
bundled SDL2 to use the system ones that have been modified to not use
`-mwindows`. Also set's `-e` to stop the script on an error (as opposed
to packaging nothing).
Uses LLVM's linker for linking yuzu. Adds -femulated-tls due to a
libstdc++ incompatibility between GCC and Clang in vulkan_common.
