bug-gnu-emacs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#74413: [PATCH] Allow to store and read repository information of VCS


From: Eli Zaretskii
Subject: bug#74413: [PATCH] Allow to store and read repository information of VCS builds
Date: Tue, 19 Nov 2024 19:08:42 +0200

> From: Björn Bidar <bjorn.bidar@thaodan.de>
> Cc: stefankangas@gmail.com,  luangruo@yahoo.com,  74413@debbugs.gnu.org
> Date: Tue, 19 Nov 2024 18:13:14 +0200
> 
> >> If the source is generated by the CI it can also store this information
> >> in the build source which then can be extracted from the ci metadata to
> >> the Emacs sources on the builder.
> >> 
> >> I can be sure that Emacs was built from that revision as much as I can
> >> trust the CI to use the sources I told it to use. If I can't trust one,
> >> I can't trust the other.
> >
> > But CI builds from Git, doesn't it?  If so, the Emacs it produces
> > already records the revision and the branch.
> 
> The source generator, the source service runs git but the builder
> doesn't.
> Because the builder doesn't have to deal with git the rebuild chain is
> smaller, build dependency changes can trigger rebuilds, and the worker
> doesn't have to have git installed.
> 
> In most cases it's not required or wanted that worker have git as the
> only allowed purpose would be to get access already existing metadata but
> not change anything on the sources itself.
> 
> Providing the metadata to the VCS emacs-repository-branch and
> emacs-repository-version with the previously generated metadata that
> belongs to the package sources removes the need for git in the build worker.

Thanks, but I still don't understand the problem you are trying to
solve.  I guess I'm missing something very fundamental here.  You are
talking about "builder", "source generator", "metadata", etc., and I
don't understand these terms.  The last sentence seems to confirm my
guess: you tage the Git revision from a repository, and then build
from another directory, which can easily cause a mismatch.  So this
feature looks completely redundant to me, based on the little I do
understand in these matters.  But I will now bow out of this
discussion; if Stefan (and others) think it's okay, I won't object.

Thank you for your patience.





reply via email to

[Prev in Thread] Current Thread [Next in Thread]