-
-
Notifications
You must be signed in to change notification settings - Fork 3.7k
tree-wide: fix builds of some packages #26655
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
tree-wide: fix builds of some packages #26655
Conversation
47dbe60 to
218eb7b
Compare
218eb7b to
fbef8de
Compare
|
Biswa96's patch is identical to the one I had here. |
fbef8de to
e8c9e5e
Compare
packages/ccls/build.sh
Outdated
| _COMMIT_DATE=2025.08.15 | ||
| TERMUX_PKG_VERSION=0.${_COMMIT_DATE//./} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Mmmh I see.
ccls trips over this since it doesn't directly update $TERMUX_PKG_VERSION so the diff check doesn't detect a version change.
We've mostly stopped using global $_COMMIT_DATE or similar helper variables.
I think putting the commit date into the version would be the easier solution.
I'll put a potential fix for the termux_step_pre_configure() check below.
| _COMMIT_DATE=2025.08.15 | |
| TERMUX_PKG_VERSION=0.${_COMMIT_DATE//./} | |
| TERMUX_PKG_VERSION=0.20250815 |
Updating lint-packages.sh to evaluate the build script would add more complexity to the version check.
ccls isn't the only package affected by this either.
rg -l 'TERMUX_PKG_VERSION=.*\$' {,root-,x11-}packages shows 52 other matches.
packages/aapt/build.sh
packages/angle-android/build.sh
packages/bash/build.sh
packages/c-toxcore/build.sh
packages/cloneit/build.sh
packages/ecj/build.sh
packages/enblend/build.sh
packages/ffmpegthumbnailer/build.sh
packages/flang/build.sh
packages/gn/build.sh
packages/gnushogi/build.sh
packages/graphene/build.sh
packages/influxdb/build.sh
packages/libcdk/build.sh
packages/libchromaprint/build.sh
packages/libgxps/build.sh
packages/libimobiledevice/build.sh
packages/libllvm/build.sh
packages/mktorrent/build.sh
packages/ovmf/build.sh
packages/pass-otp/build.sh
packages/povray/build.sh
packages/python2/build.sh
packages/quickjs/build.sh
packages/readline/build.sh
packages/rpm/build.sh
packages/simh/build.sh
packages/simulavr/build.sh
packages/subtitleripper/build.sh
packages/swiftshader/build.sh
packages/sysprop/build.sh
packages/tcc/build.sh
packages/telegram-bot-api/build.sh
root-packages/encfs/build.sh
root-packages/frida/build.sh
root-packages/open-zwave/build.sh
root-packages/usbmuxd/build.sh
x11-packages/clutter-gst/build.sh
x11-packages/clutter-gtk/build.sh
x11-packages/clutter/build.sh
x11-packages/cogl/build.sh
x11-packages/gtk2/build.sh
x11-packages/gtkmm2/build.sh
x11-packages/libdbusmenu/build.sh
x11-packages/libhandy/build.sh
x11-packages/libxml++-2.6/build.sh
x11-packages/sdl-image/build.sh
x11-packages/sdl-mixer/build.sh
x11-packages/sdl-net/build.sh
x11-packages/sdl-ttf/build.sh
x11-packages/tinyemu/build.sh
x11-packages/xfce4-datetime-plugin/build.shI'll need to look into those to decide if it's worth the extra complexity in the linter, or if fixing it in the build scripts would be easier.
_COMMIT should probably stay up here as long as we're build off a non-tag version.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good news, git show actually supports showing files based on file path, which I didn't know about.
So I can just use some more accursed rev syntax1 to essentially cat the build.sh at the fork point of the branch.
Footnotes
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good news, git show actually supports showing files based on file path, which I didn't know about. So I can just use some more accursed rev syntax1 to essentially cat the build.sh at the fork point of the branch.
Footnotes
- termux@0970a2f patches Termux's `time.h`, but unfortunately, that patch is not compatible with the `-D__ANDROID_UNAVAILABLE_SYMBOLS_ARE_WEAK__` argument internally used by `cmus`, so copying the vanilla `time.h` from its backup folder and using it specifically for only `cmus` is necessary.
- Also unvendor `libtinyxml2` to fix build error in the outdated vendored `libtinyxml2`
- Also fix the CMake header installation path to prevent `ERROR: Package contains directory "$PREFIX/$PREFIX"`
…ng is wrong with the original `TERMUX_PKG_SRCURL` - It's hard to describe what happened to the original download URL except maybe "it became a `.html` file filled with random symbols" https://galculator.mnim.org/downloads.html
…ained and fails to build - Cherry-picks patriciogonzalezvivo/vera@b6e3e12 , which disabled the `exiv2` dependency in upstream's development branch - Fixes cross-build pollution `exiv2` -> `glslviewer`
e8c9e5e to
fbbcf9c
Compare
|
I will merge it in 24 hours if no other problems are found. |
Progress on Some packages cannot be built 3: Revenge of the Xiph #23492
biboumi: fix build with CMake 4ccls: bump to fix fetching stale git submodules of the sourcecmus: fix build with NDK r28cencfs: fix build with CMake 4galculator: switch to official GitHub mirror because something weird happened to the original source URLglslviewer: disable unmaintained dependency onexiv2libunqlite: fix build with CMake 4python-cmake: updateTERMUX_PKG_SHA256quick-lint-js: fix build with CMake 4rapidjson: fix build with CMake 4tex-gyre: update license filemilkytracker: fix build with CMake 4xpdf: fix build with CMake 4