-
-
Notifications
You must be signed in to change notification settings - Fork 6.4k
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
No way to tell what code is in the daily snapshot #14363
Comments
Would adding the short HEAD commit hash to the filename work? Like : |
That would work, but it means that the daily snapshot name is no longer
predictable. The autocurl script at https://curl.se/dev/howto.html would
break and any builds people are running based on it, the autobuilds I used to
run would also have broken, as would anyone else doing snapshot builds but
assuming the name. Test Clutch uses the trick of searching for the "NEWEST"
comment at https://curl.se/snapshots/ so it should survive, unless the
directory name inside the tar ball also changes to include the hash which would
break it.
All these things can be fixed, so I'm fine with it if you think that's the best
way forward.
|
The latest snapshot is not truly predictable now either since there is no telling exactly when there is a new snapshot to download - which has shown the times where the daily snapshot builds have failed . That's why we offer the
Hm, yes right. I suppose that approach might still be used... |
How about this. A generated
I'll make a PR with the necessary changes and then we can make the snapshots done using this option. It does break the reproducibility, but we would only use this for daily snapshots and not for actual releases. |
If the second argument to the script is "commit", then this will generate a file named `docs/snapshot-commit.txt` that contains the latest commit hash (git rev-parse --short HEAD) at the time the script runs. Doing this breaks the reproducibility so it will not be used for "real" releases but is meant for automated daily snapshots and similar. Reported-by: Dan Fandrich Fixes #14363
That sounds fine to me. Having "snapshot-" in the name seems unnecessary; should
this ever be added to a regular release, it would need a new name. And it
might as well be a full hash instead of the short version.
|
If the second argument to the script is "commit", then this will generate a file named `docs/snapshot-commit.txt` that contains the latest commit hash (git rev-parse --short HEAD) at the time the script runs. Doing this breaks the reproducibility so it will not be used for "real" releases but is meant for automated daily snapshots and similar. Reported-by: Dan Fandrich Fixes #14363
If the second argument to the script is "commit", then this will generate a file named `docs/tarball-commit.txt` that contains the latest commit hash (git rev-parse HEAD) at the time the script runs. Doing this breaks the reproducibility so it will not be used for "real" releases but is meant for automated daily snapshots and similar. Reported-by: Dan Fandrich Fixes #14363 Closes #14369
This disambiguates the source code being tested. The output format is the same as when testing out of a git repo, but with no description and a long hash. Ref: #14363 Closes #X
I did this
Until PR #14331, the daily snapshot contained the file
CHANGES
that contained the git log of the code in the tar ball. Since that has been removed, there is no longer any way to tell what code is in the daily snapshot. The date itself lets you make an approximate guess, but it can't get you closer than about 7 commits (on average).I expected the following
There should be a way to tell exactly which commit the snapshot represents to replace
CHANGES
. Some possibilities:This is an issue for Test Clutch, since there is now no way to associate test results from a snapshot build to the code that generated those results.
curl/libcurl version
8.10.0-DEV (commit 8a9c227)
operating system
OS independent
The text was updated successfully, but these errors were encountered: