Re: [PATCH 07/18] link_alt_odb_entry: handle normalize_path errors

On Mon, Nov 07, 2016 at 05:12:43PM -0800, Bryan Turner wrote:

Actually, I think we want to keep normalizing, as it is possible for relative paths to normalize correctly (e.g., "foo/../bar"). We just need to ignore the error, which is easy.

The patch is below, and is the absolute minimum I think we'd need for v2.11.

Beyond that, we could go further:

  a. Actually make a real absolute path based on getcwd(), which would
     protect against later chdir() calls, and possibly help with
     duplicate suppression. I'm not sure there are actually any
     triggerable bugs here, so I went for the minimal fix.

  b. Pick a more sane base for the absolute path, like $GIT_DIR. If we
     did so, then people using relative paths in
     GIT_ALTERNATE_OBJECT_DIRECTORIES from a bare repo would continue to
     work, and people in non-bare repositories would have to add an
     extra ".." to most of their paths. So a slight regression, but
     saner overall semantics.

     Making it relative to the object directory ($GIT_DIR/objects, or
     even whatever is in $GIT_OBJECT_DIRECTORY) makes even more sense
     to me, but would regress even the bare case (and would probably be
     "interesting" along with the tmp-objdir stuff, which sets
     $GIT_OBJECT_DIRECTORY on the fly, as that would invalidate
     $GIT_ALTERNATE_OBJECT_DIRECTORIES).

I'm inclined to leave those to anybody interested post-v2.11 (or never, if nobody cares). But it would be pretty trivial to do (a) as part of this initial fix if anybody feels strongly.

The patch at the end of his mail obviously passes the newly-added tests for me, but please confirm that it fixes your test suite.

I gather your suite is about noticing behavior changes between different versions. For cases where we know there is an obvious right behavior, it would be nice if you could contribute them as patches to git's test suite. This case was overlooked because there was no test coverage at all.

Barring that, running your suite and giving easily-reproducible problem reports is valuable. The earlier the better. So I am happy to see this on -rc0, and not on the final release. Periodically running it on "master" during the development cycle would have caught it even sooner.

I would be happy for somebody to pick that up, too. It has been powering GitHub's tree-view for several years now, but I know there are some rough edges as well as opportunities to optimize it.

-- >8 --
Subject: [PATCH] alternates: re-allow relative paths from environment

Commit 670c359da (link_alt_odb_entry: handle normalize_path errors, 2016-10-03) regressed the handling of relative paths in the GIT_ALTERNATE_OBJECT_DIRECTORIES variable. It's not entirely clear this was ever meant to work, but it _has_ worked for several years, so this commit restores the original behavior.

When we get a path in GIT_ALTERNATE_OBJECT_DIRECTORIES, we add it the path to the list of alternate object directories as if it were found in objects/info/alternates, but with one difference: we do not provide the link_alt_odb_entry() function with a base for relative paths. That function doesn't turn it into an absolute path, and we end up feeding the relative path to the strbuf_normalize_path() function.

Most relative paths break out of the top-level directory (e.g., "../foo.git/objects"), and thus normalizing fails. Prior to 670c359da, we simply ignored the error, and due to the way normalize_path_copy() was implemented it happened to return the original path in this case. We then accessed the alternate objects using this relative path.

By storing the relative path in the alt_odb list, the path is relative to wherever we happen to be at the time we do an object lookup. That means we look from $GIT_DIR in a bare repository, and from the top of the worktree in a non-bare repository.

If this were being designed from scratch, it would make sense to pick a stable location (probably $GIT_DIR, or even the object directory) and use that as the relative base, turning the result into an absolute path. However, given the history, at this point the minimal fix is to match the pre-670c359da behavior.

We can do this simply by ignoring the error when we have no relative base and using the original value (which we now reliably have, thanks to strbuf_normalize_path()).

That still leaves us with a relative path that foils our duplicate detection, and may act strangely if we ever chdir() later in the process. We could solve that by storing an absolute path based on getcwd(). That may be a good future direction; for now we'll do just the minimum to fix the regression.

The new t5615 script demonstrates the fix in its final three tests. Since we didn't have any tests of the alternates environment variable at all, it also adds some tests of absolute paths.

Signed-off-by: Jeff King <peff@peff.net>
---
 sha1_file.c              |  2 +-
 t/t5615-alternate-env.sh | 71 ++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 72 insertions(+), 1 deletion(-)
 create mode 100755 t/t5615-alternate-env.sh