From: Junio C Hamano <gitster@pobox•com>
To: Linus Torvalds <torvalds@linux-foundation•org>
Cc: Git Mailing List <git@vger•kernel.org>,
Frank <streamlake@tiscali•it>,
Dmitry Potapov <dpotapov@gmail•com>
Subject: Re: [PATCH 7/7] Make unpack-tree update removed files before any updated files
Date: Sat, 22 Mar 2008 22:49:53 -0700 [thread overview]
Message-ID: <7vbq56f0qm.fsf@gitster.siamese.dyndns.org> (raw)
In-Reply-To: alpine.LFD.1.00.0803221038320.3020@woody.linux-foundation.org
Linus Torvalds <torvalds@linux-foundation•org> writes:
> Ok, this one looks - and is, really - trivial, but it's actually the only
> one in the whole series that I'm even remotely nervous about. First off,
> it actually does what it does regardless of that "core.ignorecase"
> variable, but that wouldn't worry me if it wasn't for the fact that I
> don't remember/understand what the heck that "last_symlink" logic was
> there for.
last_symlink is just a cached information used by underlying
has_symlink_leading_path() function for optimization.
The motivation behind has_symlink_leading_path() is reasonably well
described in
- f859c84: Add has_symlink_leading_path() function., 2007-05-11,
- 64cab59: apply: do not get confused by symlinks in the middle,
2007-05-11, and
- 16a4c61: read-tree -m -u: avoid getting confused by intermediate
symlinks., 2007-05-10
The short version is that:
- sometimes we want to make sure a path a/b/c/d exists (or does not
exist) in the work tree;
- however, !lstat("a/b/c/d") is not quite it. if a/b is a symlink in the
work tree that points at somewhere that happens to have c/d underneath,
!lstat() says "yeah, there is", but that one is _different_ from what
checking out a cache entry a/b/c/d would produce (because in that case
we will remove a/b symlink, create a/b/ directory and deposit blob b
there).
- So we often need to see if a given path has symlink component in the
leading part in the work tree (e.g. given "a/b/c/d", we would need to
check if any of "a", "a/b", "a/b/c" is a symlink).
The function has_symlink_leading_path() answers that question, and its
second argument is a buffer to cache "the last work tree path found to be
a symlink", so if you call it with "a/b/c/d" and then with "a/b/c/e" in
the above example situation, the second call can re-use the information
the first call found out, which is "a/b is a symlink".
I do not think your patch breaks the passing around of last_symlink cached
information. Although the three commits I quoted above are all backed by
real-world breakage cases that they did fix, the issues they deal with are
indeed tricky cases. Although your patch (the change in 7/7) should not
make any difference to the issues, thinking about them is already making
me feel nervous.
next prev parent reply other threads:[~2008-03-23 5:50 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-03-22 17:21 [PATCH 0/7] Case-insensitive filesystem support, take 1 Linus Torvalds
2008-03-22 17:22 ` [PATCH 1/7] Make unpack_trees_options bit flags actual bitfields Linus Torvalds
2008-03-22 17:25 ` [PATCH 2/7] Move name hashing functions into a file of its own Linus Torvalds
2008-03-22 17:28 ` [PATCH 3/7] Make "index_name_exists()" return the cache_entry it found Linus Torvalds
2008-03-22 17:30 ` [PATCH 4/7] Make hash_name_lookup able to do case-independent lookups Linus Torvalds
2008-03-22 17:33 ` [PATCH 5/7] Add 'core.ignorecase' option Linus Torvalds
2008-03-22 17:38 ` [PATCH 6/7] Make branch merging aware of underlying case-insensitive filsystems Linus Torvalds
2008-03-22 17:45 ` [PATCH 7/7] Make unpack-tree update removed files before any updated files Linus Torvalds
2008-03-22 18:06 ` [PATCH 0/7] Final words Linus Torvalds
2008-03-22 18:28 ` Linus Torvalds
2008-03-22 21:19 ` [PATCH 8/7] When adding files to the index, add support for case-independent matches Linus Torvalds
2008-03-22 21:22 ` [PATCH 9/7] Make git-add behave more sensibly in a case-insensitive environment Linus Torvalds
2008-03-23 5:49 ` Junio C Hamano [this message]
2008-03-23 6:13 ` [PATCH 6/7] Make branch merging aware of underlying case-insensitive filsystems Junio C Hamano
2008-03-23 15:41 ` Linus Torvalds
2008-03-22 17:36 ` [PATCH 1/7] Make unpack_trees_options bit flags actual bitfields Johannes Schindelin
2008-03-22 17:47 ` Linus Torvalds
2008-03-22 17:57 ` Johannes Schindelin
2008-03-22 22:01 ` [PATCH] t0050: Set core.ignorecase case to activate case insensitivity Steffen Prohaska
2008-03-25 6:57 ` [PATCH] git-init: autodetect core.ignorecase Dmitry Potapov
2008-03-25 9:59 ` Johannes Schindelin
2008-03-25 10:49 ` [PATCH v2] " Dmitry Potapov
2008-03-25 11:03 ` [PATCH] " Dmitry Potapov
2008-03-25 8:14 ` [PATCH 0/7] Case-insensitive filesystem support, take 1 Dmitry Potapov
2008-03-25 21:04 ` Linus Torvalds
2008-03-26 2:46 ` Dmitry Potapov
2008-03-26 3:37 ` Linus Torvalds
2008-03-25 11:39 ` Derek Fawcus
2008-03-25 18:26 ` Jan Hudec
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=7vbq56f0qm.fsf@gitster.siamese.dyndns.org \
--to=gitster@pobox$(echo .)com \
--cc=dpotapov@gmail$(echo .)com \
--cc=git@vger$(echo .)kernel.org \
--cc=streamlake@tiscali$(echo .)it \
--cc=torvalds@linux-foundation$(echo .)org \
/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