public inbox for git@vger.kernel.org 
 help / color / mirror / Atom feed
From: Felipe Contreras <felipe.contreras@gmail•com>
To: Todd Zullinger <tmz@pobox•com>,
	Felipe Contreras <felipe.contreras@gmail•com>
Cc: git@vger•kernel.org, Jeff King <peff@peff•net>,
	Junio C Hamano <gitster@pobox•com>
Subject: Re: [PATCH 06/18] version-gen: remove redundant check
Date: Fri, 14 Apr 2023 13:32:43 -0600	[thread overview]
Message-ID: <6439aa5b2c776_6058729489@chronos.notmuch> (raw)
In-Reply-To: <ZDmjIqNwVMESgHnn@pobox.com>

Todd Zullinger wrote:
> Felipe Contreras wrote:
> > Todd Zullinger wrote:
> >> Felipe Contreras wrote:
> >>> If we are not in a git repository `git describe` will fail anyway.
> >> 
> >> The parent directory may be a git repository though.  The
> >> current code ensures that we're running `git describe` in
> >> the proper repository.
> > 
> > How exactly does it do that?
> > 
> > The current code expects the cwd to be the git repo, run it in any other
> > directory and it will generate GIT-VERSION-FILE in that directory, which is
> > clearly not intended.
> 
> Whether it's fool proof isn't really my point.  It did
> attempt to check that .git was a file or directory.  Not
> checking at all isn't necessarily an improvement, was my
> concern.
> 
> >> If we drop this, aren't we breaking things for someone
> >> building a git tarball which is in a subdirectory of a git
> >> repository?
> > 
> > How exactly would this hypothetical person build such a tarball?
> > 
> >   git init /tmp/foo
> >   mkdir -p /tmp/foo/bar
> >   cd /tmp/foo/bar
> >   make -C ~/dev/git dist
> 
> If I have a git repo, say ~/fedora/git which contains the
> fedora packaging (spec file, etc.) and extract a git archive
> in this directory, the describe will now pick up the data
> from the parent git directory, won't it?
> 
>     $ git -C ~/fedora clone https://src.fedoraproject.org/rpms/git.git
>     $ cd ~/fedora/git
>     $ git -C ~/upstream/git archive --format=tar --prefix=git/ HEAD | tar xf -
>     $ cd git
>     $ make GIT-VERSION-GEN
>     $ cat GIT-VERSION-FILE 
>     GIT_VERSION = 

I don't think this is a realistic use-case, but supposing it is, what would be
the desired outcome in this case?

 GIT_VERSION = 2.40.GIT

?

> The version file in the tarballs prevents this from
> happening in the most common case, but it still feels like
> this is loosening things a little more than it should.

If we care about this, the same behavior can be achieved with GIT_CEILING_DIRECTORIES:

 GIT_CEILING_DIRECTORIES=$(cd .. && pwd) git describe ...

-- 
Felipe Contreras

  reply	other threads:[~2023-04-14 19:32 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-04-14 12:18 [PATCH 00/18] version-gen: complete revamp/rewrite Felipe Contreras
2023-04-14 12:18 ` [PATCH 01/18] version-gen: reorganize Felipe Contreras
2023-04-14 12:18 ` [PATCH 02/18] version-gen: trivial cleanup Felipe Contreras
2023-04-14 12:18 ` [PATCH 03/18] version-gen: refactor default version Felipe Contreras
2023-04-14 12:18 ` [PATCH 04/18] version-gen: simplify v prefix removal Felipe Contreras
2023-04-14 12:18 ` [PATCH 05/18] version-gen: simplify update check Felipe Contreras
2023-04-14 12:18 ` [PATCH 06/18] version-gen: remove redundant check Felipe Contreras
2023-04-14 15:11   ` Todd Zullinger
2023-04-14 17:47     ` Felipe Contreras
2023-04-14 19:01       ` Todd Zullinger
2023-04-14 19:32         ` Felipe Contreras [this message]
2023-04-14 19:41           ` Todd Zullinger
2023-04-14 12:18 ` [PATCH 07/18] version-gen: simplify `git describe` checks Felipe Contreras
2023-04-14 12:18 ` [PATCH 08/18] version-gen: simplify dirty check Felipe Contreras
2023-04-14 12:18 ` [PATCH 09/18] version-gen: move describe fix into function Felipe Contreras
2023-04-14 12:18 ` [PATCH 10/18] version-gen: describe and sed in one go Felipe Contreras
2023-04-14 12:18 ` [PATCH 11/18] version-gen: refactor describe function Felipe Contreras
2023-04-14 12:18 ` [PATCH 12/18] version-gen: do v fix only when necessary Felipe Contreras
2023-04-14 12:18 ` [PATCH 13/18] version-gen: move v fix into sed Felipe Contreras
2023-04-14 12:18 ` [PATCH 14/18] version-gen: refactor main functionality Felipe Contreras
2023-04-14 12:18 ` [PATCH 15/18] version-gen: remove default version Felipe Contreras
2023-04-14 12:18 ` [PATCH 16/18] version-gen: refactor GIT_VERSION string Felipe Contreras
2023-04-14 12:18 ` [PATCH 17/18] version-gen: get rid of GVF variable Felipe Contreras
2023-04-14 12:18 ` [PATCH 18/18] version-gen: generate proper interim versions Felipe Contreras

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=6439aa5b2c776_6058729489@chronos.notmuch \
    --to=felipe.contreras@gmail$(echo .)com \
    --cc=git@vger$(echo .)kernel.org \
    --cc=gitster@pobox$(echo .)com \
    --cc=peff@peff$(echo .)net \
    --cc=tmz@pobox$(echo .)com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox