Publish individual bins for Linux-arm64 with regression test for cross-build#10606
Publish individual bins for Linux-arm64 with regression test for cross-build#10606oech3 wants to merge 1 commit intouutils:mainfrom
Conversation
|
GNU testsuite comparison: |
This comment was marked as resolved.
This comment was marked as resolved.
|
@oech3 I'll take a look this weekend. I have no time today. |
|
ok. thankyou |
9742559 to
8cfa4ba
Compare
|
GNU testsuite comparison: |
|
GNU testsuite comparison: |
443d211 to
290ff05
Compare
|
If we publish bins, it should be release profile instead of release-small. It made diff larger... |
|
GNU testsuite comparison: |
|
GNU testsuite comparison: |
CodSpeed Performance ReportMerging this PR will degrade performance by 3.62%Comparing Summary
Performance Changes
Footnotes
|
Ecordonnier
left a comment
There was a problem hiding this comment.
I've added some comments inline. Please also extend the PR description.
| test -f /tmp/usr/local/bin/uu-tty | ||
| test -f /tmp/usr/local/libexec/uu-coreutils/libstdbuf.* | ||
| sudo apt-get install -y gcc-aarch64-linux-gnu | ||
| RUSTFLAGS="${RUSTFLAGS} -C strip=symbols" |
There was a problem hiding this comment.
| RUSTFLAGS="${RUSTFLAGS} -C strip=symbols" | |
| export RUSTFLAGS="${RUSTFLAGS} -C strip=symbols" |
There was a problem hiding this comment.
shellcheck complains if we don't split export. (done at L329)
| test -f /tmp/usr/local/libexec/uu-coreutils/libstdbuf.* | ||
| sudo apt-get install -y gcc-aarch64-linux-gnu | ||
| RUSTFLAGS="${RUSTFLAGS} -C strip=symbols" | ||
| export RUSTFLAGS |
There was a problem hiding this comment.
| export RUSTFLAGS |
| build_makefile: | ||
| name: Build/Makefile | ||
| needs: [ min_version, deps ] | ||
| permissions: |
There was a problem hiding this comment.
As far as I understand the "build_makefile" jobs are test jobs. The "build" job at line 627 already has "permissions: contents: write" and already has one line for aarch64-unknown-linux-gnu.
Are you trying to publish files from the build_makefile job instead of from the build job so that you can get libstdbuf.so compiled with feat_external_stdbuf? Or why not use the build job?
There was a problem hiding this comment.
Should it not rather be something like this in the publish job around line 534?
matrix:
job:
- { os: ubuntu-latest, target: x86_64-unknown-linux-gnu, features: feat_os_unix }
- { os: ubuntu-latest, target: aarch64-unknown-linux-gnu, features: feat_os_unix_gnueabihf, use-cross: true }
There was a problem hiding this comment.
x64 individual bins are published by reusing compute_size:. libstdbuf.so is also one of a reason to publish from here.
Ofcause I'm considering to merge jobs for publishing (and deduplicating release build). But currently, I can't.
There was a problem hiding this comment.
(We should stop/merge duplicated build for compute_size:)
| ! test -f /tmp/usr/local/share/zsh/site-functions/_install | ||
| ! test -f /tmp/usr/local/share/bash-completion/completions/head.bash | ||
| ! test -f /tmp/usr/local/share/fish/vendor_completions.d/cat.fish | ||
| # We publish them instead of discarding |
There was a problem hiding this comment.
I think that in order to change what gets released it would be good to have approval from @sylvestre . Can you please open an issue regarding the addition of individual arm binaries to the releases to discuss this?
There was a problem hiding this comment.
Actually, I just don't want to discard bins with release build (not interested in arm).
Cross-build for FreeBSD might better since Linux-arm coreutils is already published.
There was a problem hiding this comment.
As far as I know, 99,999% of users will get uutils coreutils using their package-manager, and the distros like freebsd will use the source code of uutils coreutils and compile it themselves and not use the binary packages. So I am not sure it is useful to publish individual binaries for arm or freebsd? What is the use case? That is why I suggested to create an issue to discuss it.
There was a problem hiding this comment.
Package manager does not provide binaries from main branch and most distributor provides single binary.
943d6fc to
9f9c52b
Compare
|
GNU testsuite comparison: |
@Ecordonnier