git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Matheus Tavares Bernardino <matheus.bernardino@usp.br>
To: Thomas Gummerer <t.gummerer@gmail.com>
Cc: git <git@vger.kernel.org>,
	"Ævar Arnfjörð Bjarmason" <avarab@gmail.com>,
	"Christian Couder" <christian.couder@gmail.com>,
	"Nguyễn Thái Ngọc Duy" <pclouds@gmail.com>,
	"Kernel USP" <kernel-usp@googlegroups.com>,
	"Benoit Pierre" <benoit.pierre@gmail.com>,
	"Junio C Hamano" <gitster@pobox.com>,
	"Johannes Schindelin" <Johannes.Schindelin@gmx.de>
Subject: Re: [GSoC][PATCH v4 2/7] clone: better handle symlinked files at .git/objects/
Date: Fri, 29 Mar 2019 11:27:54 -0300	[thread overview]
Message-ID: <CAHd-oW78Cwhq8sFFjOX4c6kS8JsEUUd_RGgheJhdN6MLAr+G2Q@mail.gmail.com> (raw)
In-Reply-To: <20190328221049.GK32487@hank.intra.tgummerer.com>

On Thu, Mar 28, 2019 at 7:10 PM Thomas Gummerer <t.gummerer@gmail.com> wrote:
>
> On 03/22, Matheus Tavares wrote:
> > There is currently an odd behaviour when locally clonning a repository
> > with symlinks at .git/objects: using --no-hardlinks all symlinks are
> > dereferenced but without it Git will try to hardlink the files with the
> > link() function, which has an OS-specific behaviour on symlinks. On OSX
> > and NetBSD, it creates a hardlink to the file pointed by the symlink
> > whilst on GNU/Linux, it creates a hardlink to the symlink itself.
> >
> > On Manjaro GNU/Linux:
> >     $ touch a
> >     $ ln -s a b
> >     $ link b c
> >     $ ls -li a b c
> >     155 [...] a
> >     156 [...] b -> a
> >     156 [...] c -> a
> >
> > But on NetBSD:
> >     $ ls -li a b c
> >     2609160 [...] a
> >     2609164 [...] b -> a
> >     2609160 [...] c
> >
> > It's not good to have the result of a local clone to be OS-dependent and
> > since the behaviour on GNU/Linux may result in broken symlinks, let's
> > re-implement it with linkat() instead of link() using a flag to always
> > follow symlinks and make the hardlink be to the pointed file. With this,
> > besides standardizing the behaviour, no broken symlinks will be
> > produced. Also, add tests for symlinked files at .git/objects/.
> >
> > Note: Git won't create symlinks at .git/objects itself, but it's better
> > to handle this case and be friendly with users who manually create them.
> >
> > Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
> > Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
> > Co-authored-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
> > ---
> >  builtin/clone.c            |  2 +-
> >  t/t5604-clone-reference.sh | 26 +++++++++++++++++++-------
> >  2 files changed, 20 insertions(+), 8 deletions(-)
> >
> > diff --git a/builtin/clone.c b/builtin/clone.c
> > index 50bde99618..b76f33c635 100644
> > --- a/builtin/clone.c
> > +++ b/builtin/clone.c
> > @@ -443,7 +443,7 @@ static void copy_or_link_directory(struct strbuf *src, struct strbuf *dest,
> >               if (unlink(dest->buf) && errno != ENOENT)
> >                       die_errno(_("failed to unlink '%s'"), dest->buf);
> >               if (!option_no_hardlinks) {
> > -                     if (!link(src->buf, dest->buf))
> > +                     if (!linkat(AT_FDCWD, src->buf, AT_FDCWD, dest->buf, AT_SYMLINK_FOLLOW))
>
> This line is starting to get a bit long, might be worth breaking it up
> to keep to 80 characters per line.
>
> I notice that we are currently not using 'linkat()' anywhere else in
> our codebase.  It looks like it has been introduced in POSIX.1-2008,
> which sounds fairly recent by git's standards.  So I wonder if this is
> really supported on all platforms that git is being built on.
>
> I also wonder what would need to be done on Windows if we were to
> introduce this.  I see we define the 'link()' function in
> 'compat/mingw.c' for that currently, so I guess something similar
> would be needed for 'linkat()'.  I added Dscho to Cc for Windows
> expertise.

Ok, what if instead of using linkat() we use 'realpath(const char
*path, char *resolved_path)', which will resolve any symlinks at
'path' and store the canonical path at 'resolved_path'? Then, we can
still keep using link() but now, with the certainty that all platforms
will have a consistent behaviour? (also, realpath() is POSIX.1-2001)
Would that be a better idea?

> While I agree with the goal of consistency accross all platforms here,
> I don't know if it's actually worth going through the pain of doing
> that, especially for somewhat of an edge case in local clones.
>
> If the test in the previous patch passes on all platforms, I'd be okay
> with just calling the behaviour here undefined, especially as git
> would never actually create symlinks in the .git/objects directory.

  parent reply	other threads:[~2019-03-29 14:28 UTC|newest]

Thread overview: 127+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-02-26  5:17 [WIP RFC PATCH v2 0/5] clone: dir iterator refactoring with tests Matheus Tavares
2019-02-26  5:18 ` [WIP RFC PATCH v2 1/5] dir-iterator: add flags parameter to dir_iterator_begin Matheus Tavares
2019-02-26 12:01   ` Duy Nguyen
2019-02-27 13:59     ` Matheus Tavares Bernardino
2019-02-26  5:18 ` [WIP RFC PATCH v2 2/5] clone: test for our behavior on odd objects/* content Matheus Tavares
2019-02-26  5:18 ` [WIP RFC PATCH v2 3/5] clone: copy hidden paths at local clone Matheus Tavares
2019-02-26 12:13   ` Duy Nguyen
2019-02-26  5:18 ` [WIP RFC PATCH v2 4/5] clone: extract function from copy_or_link_directory Matheus Tavares
2019-02-26 12:18   ` Duy Nguyen
2019-02-27 17:30     ` Matheus Tavares Bernardino
2019-02-27 22:45       ` Thomas Gummerer
2019-02-27 22:50         ` Matheus Tavares Bernardino
2019-02-26  5:18 ` [WIP RFC PATCH v2 5/5] clone: use dir-iterator to avoid explicit dir traversal Matheus Tavares
2019-02-26 11:35   ` Ævar Arnfjörð Bjarmason
2019-02-26 12:32   ` Duy Nguyen
2019-02-26 12:50     ` Ævar Arnfjörð Bjarmason
2019-02-27 17:40     ` Matheus Tavares Bernardino
2019-02-28  7:13       ` Duy Nguyen
2019-02-28  7:53       ` Ævar Arnfjörð Bjarmason
2019-02-26 11:36 ` [WIP RFC PATCH v2 0/5] clone: dir iterator refactoring with tests Ævar Arnfjörð Bjarmason
2019-02-26 12:20   ` Duy Nguyen
2019-02-26 12:28 ` [RFC PATCH v3 " Ævar Arnfjörð Bjarmason
2019-02-26 20:56   ` Matheus Tavares Bernardino
2019-03-22 23:22   ` [GSoC][PATCH v4 0/7] clone: dir-iterator " Matheus Tavares
2019-03-22 23:22     ` [GSoC][PATCH v4 1/7] clone: test for our behavior on odd objects/* content Matheus Tavares
2019-03-24 18:09       ` Matheus Tavares Bernardino
2019-03-24 20:56       ` SZEDER Gábor
2019-03-26 19:43         ` Matheus Tavares Bernardino
2019-03-28 21:49       ` Thomas Gummerer
2019-03-29 14:06         ` Matheus Tavares Bernardino
2019-03-29 19:31           ` Thomas Gummerer
2019-03-29 19:42             ` SZEDER Gábor
2019-03-30  2:49             ` Matheus Tavares Bernardino
2019-03-22 23:22     ` [GSoC][PATCH v4 2/7] clone: better handle symlinked files at .git/objects/ Matheus Tavares
2019-03-28 22:10       ` Thomas Gummerer
2019-03-29  8:38         ` Ævar Arnfjörð Bjarmason
2019-03-29 20:15           ` Thomas Gummerer
2019-03-29 14:27         ` Matheus Tavares Bernardino [this message]
2019-03-29 20:05           ` Thomas Gummerer
2019-03-30  5:32             ` Matheus Tavares Bernardino
2019-03-30 19:27               ` Thomas Gummerer
2019-04-01  3:56                 ` Matheus Tavares Bernardino
2019-03-29 15:40         ` Johannes Schindelin
2019-03-22 23:22     ` [GSoC][PATCH v4 3/7] dir-iterator: add flags parameter to dir_iterator_begin Matheus Tavares
2019-03-28 22:19       ` Thomas Gummerer
2019-03-29 13:16         ` Matheus Tavares Bernardino
2019-03-22 23:22     ` [GSoC][PATCH v4 4/7] clone: copy hidden paths at local clone Matheus Tavares
2019-03-22 23:22     ` [GSoC][PATCH v4 5/7] clone: extract function from copy_or_link_directory Matheus Tavares
2019-03-22 23:22     ` [GSoC][PATCH v4 6/7] clone: use dir-iterator to avoid explicit dir traversal Matheus Tavares
2019-03-22 23:22     ` [GSoC][PATCH v4 7/7] clone: Replace strcmp by fspathcmp Matheus Tavares
2019-03-30 22:49     ` [GSoC][PATCH v5 0/7] clone: dir-iterator refactoring with tests Matheus Tavares
2019-03-30 22:49       ` [GSoC][PATCH v5 1/7] clone: test for our behavior on odd objects/* content Matheus Tavares
2019-03-30 22:49       ` [GSoC][PATCH v5 2/7] clone: better handle symlinked files at .git/objects/ Matheus Tavares
2019-03-31 17:40         ` Thomas Gummerer
2019-04-01  3:59           ` Matheus Tavares Bernardino
2019-03-30 22:49       ` [GSoC][PATCH v5 3/7] dir-iterator: add flags parameter to dir_iterator_begin Matheus Tavares
2019-03-31 18:12         ` Thomas Gummerer
2019-04-10 20:24           ` Matheus Tavares Bernardino
2019-04-11 21:09             ` Thomas Gummerer
2019-04-23 17:07               ` Matheus Tavares Bernardino
2019-04-24 18:36                 ` Thomas Gummerer
2019-04-26  4:13                   ` Matheus Tavares Bernardino
2019-03-30 22:49       ` [GSoC][PATCH v5 4/7] clone: copy hidden paths at local clone Matheus Tavares
2019-03-30 22:49       ` [GSoC][PATCH v5 5/7] clone: extract function from copy_or_link_directory Matheus Tavares
2019-03-30 22:49       ` [GSoC][PATCH v5 6/7] clone: use dir-iterator to avoid explicit dir traversal Matheus Tavares
2019-03-30 22:49       ` [GSoC][PATCH v5 7/7] clone: replace strcmp by fspathcmp Matheus Tavares
2019-03-31 18:16       ` [GSoC][PATCH v5 0/7] clone: dir-iterator refactoring with tests Thomas Gummerer
2019-04-01 13:56         ` Matheus Tavares Bernardino
2019-05-02 14:48       ` [GSoC][PATCH v6 00/10] " Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 01/10] clone: test for our behavior on odd objects/* content Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 02/10] clone: better handle symlinked files at .git/objects/ Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 03/10] dir-iterator: add tests for dir-iterator API Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 04/10] dir-iterator: use warning_errno when possible Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 05/10] dir-iterator: refactor state machine model Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 06/10] dir-iterator: add flags parameter to dir_iterator_begin Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 07/10] clone: copy hidden paths at local clone Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 08/10] clone: extract function from copy_or_link_directory Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 09/10] clone: use dir-iterator to avoid explicit dir traversal Matheus Tavares
2019-05-02 14:48         ` [GSoC][PATCH v6 10/10] clone: replace strcmp by fspathcmp Matheus Tavares
2019-06-18 23:27         ` [GSoC][PATCH v7 00/10] clone: dir-iterator refactoring with tests Matheus Tavares
2019-06-18 23:27           ` [GSoC][PATCH v7 01/10] clone: test for our behavior on odd objects/* content Matheus Tavares
2019-06-18 23:27           ` [GSoC][PATCH v7 02/10] clone: better handle symlinked files at .git/objects/ Matheus Tavares
2019-06-18 23:27           ` [GSoC][PATCH v7 03/10] dir-iterator: add tests for dir-iterator API Matheus Tavares
2019-06-18 23:27           ` [GSoC][PATCH v7 04/10] dir-iterator: use warning_errno when possible Matheus Tavares
2019-06-18 23:27           ` [GSoC][PATCH v7 05/10] dir-iterator: refactor state machine model Matheus Tavares
2019-06-18 23:27           ` [GSoC][PATCH v7 06/10] dir-iterator: add flags parameter to dir_iterator_begin Matheus Tavares
2019-06-25 18:00             ` Junio C Hamano
2019-06-25 18:11               ` Matheus Tavares Bernardino
2019-06-26 13:34             ` Johannes Schindelin
2019-06-26 18:04               ` Junio C Hamano
2019-06-27  9:20                 ` Duy Nguyen
2019-06-27 17:23                 ` Matheus Tavares Bernardino
2019-06-27 18:48                   ` Johannes Schindelin
2019-06-27 19:33                     ` Matheus Tavares Bernardino
2019-06-28 12:51                       ` Johannes Schindelin
2019-06-28 14:16                         ` Matheus Tavares Bernardino
2019-07-01 12:15                           ` Johannes Schindelin
2019-07-03  8:57             ` SZEDER Gábor
2019-07-08 22:21               ` Matheus Tavares Bernardino
2019-06-18 23:27           ` [GSoC][PATCH v7 07/10] clone: copy hidden paths at local clone Matheus Tavares
2019-06-18 23:27           ` [GSoC][PATCH v7 08/10] clone: extract function from copy_or_link_directory Matheus Tavares
2019-06-18 23:27           ` [GSoC][PATCH v7 09/10] clone: use dir-iterator to avoid explicit dir traversal Matheus Tavares
2019-06-18 23:27           ` [GSoC][PATCH v7 10/10] clone: replace strcmp by fspathcmp Matheus Tavares
2019-06-19  4:36           ` [GSoC][PATCH v7 00/10] clone: dir-iterator refactoring with tests Matheus Tavares Bernardino
2019-06-20 20:18           ` Junio C Hamano
2019-06-21 13:41             ` Matheus Tavares Bernardino
2019-07-10 23:58           ` [GSoC][PATCH v8 " Matheus Tavares
2019-07-10 23:58             ` [GSoC][PATCH v8 01/10] clone: test for our behavior on odd objects/* content Matheus Tavares
2019-07-10 23:58             ` [GSoC][PATCH v8 02/10] clone: better handle symlinked files at .git/objects/ Matheus Tavares
2019-07-10 23:58             ` [GSoC][PATCH v8 03/10] dir-iterator: add tests for dir-iterator API Matheus Tavares
2019-07-10 23:58             ` [GSoC][PATCH v8 04/10] dir-iterator: use warning_errno when possible Matheus Tavares
2019-07-10 23:58             ` [GSoC][PATCH v8 05/10] dir-iterator: refactor state machine model Matheus Tavares
2019-07-10 23:59             ` [GSoC][PATCH v8 06/10] dir-iterator: add flags parameter to dir_iterator_begin Matheus Tavares
2019-07-10 23:59             ` [GSoC][PATCH v8 07/10] clone: copy hidden paths at local clone Matheus Tavares
2019-07-10 23:59             ` [GSoC][PATCH v8 08/10] clone: extract function from copy_or_link_directory Matheus Tavares
2019-07-10 23:59             ` [GSoC][PATCH v8 09/10] clone: use dir-iterator to avoid explicit dir traversal Matheus Tavares
2019-07-10 23:59             ` [GSoC][PATCH v8 10/10] clone: replace strcmp by fspathcmp Matheus Tavares
2019-07-11 11:56             ` [GSoC][PATCH v8 00/10] clone: dir-iterator refactoring with tests Johannes Schindelin
2019-07-11 15:24               ` Matheus Tavares Bernardino
2019-02-26 12:28 ` [RFC PATCH v3 1/5] clone: test for our behavior on odd objects/* content Ævar Arnfjörð Bjarmason
2019-02-28 21:19   ` Matheus Tavares Bernardino
2019-03-01 13:49     ` Ævar Arnfjörð Bjarmason
2019-03-13  3:17       ` Matheus Tavares
2019-02-26 12:28 ` [RFC PATCH v3 2/5] dir-iterator: add flags parameter to dir_iterator_begin Ævar Arnfjörð Bjarmason
2019-02-26 12:28 ` [RFC PATCH v3 3/5] clone: copy hidden paths at local clone Ævar Arnfjörð Bjarmason
2019-02-26 12:28 ` [RFC PATCH v3 4/5] clone: extract function from copy_or_link_directory Ævar Arnfjörð Bjarmason
2019-02-26 12:28 ` [RFC PATCH v3 5/5] clone: use dir-iterator to avoid explicit dir traversal Ævar Arnfjörð Bjarmason

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

  List information: http://vger.kernel.org/majordomo-info.html

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

  git send-email \
    --in-reply-to=CAHd-oW78Cwhq8sFFjOX4c6kS8JsEUUd_RGgheJhdN6MLAr+G2Q@mail.gmail.com \
    --to=matheus.bernardino@usp.br \
    --cc=Johannes.Schindelin@gmx.de \
    --cc=avarab@gmail.com \
    --cc=benoit.pierre@gmail.com \
    --cc=christian.couder@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=kernel-usp@googlegroups.com \
    --cc=pclouds@gmail.com \
    --cc=t.gummerer@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).