Re: [PATCH] rev-parse: Fix shell scripts whose cwd is a symlink into a git work-dir

From: Marcel M. Cary <marcel@oak.homeunix.org>
Date: Nov 25, 2008, 05:54 UTC
Message-ID: <492B9321.8090706@oak.homeunix.org>
In-Reply-To: <49290203.7060808@op5.se>

> AFAIR, it was introduced to make test-builds of really large projects in
> really deep directories with a ton of symlinks leading to the path work a
> lot faster.

Andreas,

I see value in keeping Git very fast. That is, after all, why I chose Git over Mercurial. Do you know where that discussion was, if was in the archives? I found these reasons to avoid absolute paths in the git archives:

* paths with more components are slower to work with (in the context of
  add and diff, which deal with many many paths)
* absolute paths may exceed PATH_MAX while relative ones didn't
* getcwd() will fail if parent directories are not executable, or on
  some platforms, if parent directories are not readable

My impression is that the performance issue is probably not significant for cd_to_toplevel since it's not in a tight inner loop, and dito for other potential callers of --show-cdup. The PATH_MAX seems to be a restriction elsewhere in the code already.

Even if there were a scenario that put --show-cdup in a tight loop, I wonder whether current implementation provides much performance benefit, at least when bash is the calling language: bash seems to make the relative path absolute anyway inside the "cd" builtin.

The commit (5f94c730) that introduces that code doesn't mention performance. It compares to:

  git rev-parse --show-prefix | sed -e 's|[^/][^/]*|..|g'

I also noticed that this failure case with "--show-cdup" in a symlinked directory has come up more than once before. http://marc.info/?l=git&m=122452534912000&w=2 http://marc.info/?l=git&m=121613416212958&w=2 https://kerneltrap.org/mailarchive/git/2007/4/25/244653/thread

Marcel