git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Han Xin <hanxin.hx@bytedance.com>
To: me@ttaylorr.com
Cc: chiyutianyi@gmail.com, derrickstolee@github.com,
	git@vger.kernel.org, haiyangtand@gmail.com,
	jonathantanmy@google.com, ps@pks.im,
	Han Xin <hanxin.hx@bytedance.com>
Subject: [RFC PATCH 0/2] Re: An endless loop fetching issue with partial clone, alternates and commit graph
Date: Thu, 16 Jun 2022 11:38:31 +0800	[thread overview]
Message-ID: <cover.1655350442.git.hanxin.hx@bytedance.com> (raw)
In-Reply-To: <YqlBjET0tf7V9/sg@nand.local>

On Wed, Jun 15, 2022 at 10:18 AM Taylor Blau <me@ttaylorr.com> wrote:
>
> [+cc Stolee]
>
> On Tue, Jun 14, 2022 at 03:25:13PM +0800, Haiyng Tan wrote:
> > I think it's caused by using lazy-fetch in
> > deref_without_lazy_fetch_extended().  In lookup_commit_in_graph(),
> > lazy-fetch is initiated by repo_has_object_file() used.  has_object()
> > should be used, it's no-lazy-fetch.
>
> Hmm. Are there cases where lookup_commit_in_graph() is expected to
> lazily fetch missing objects from promisor remotes? If so, then this
> wouldn't quite work. If not, then this seems like an appropriate fix to
> me.
>
> Thanks,
> Taylor

We can see the use of has_object() in RelNotes/2.29.0.txt[1]:
   * A new helper function has_object() has been introduced to make it
     easier to mark object existence checks that do and don't want to
     trigger lazy fetches, and a few such checks are converted using it.

Let's see the difference between has_object() and repo_has_object_file():
    int has_object(struct repository *r, const struct object_id *oid,
            unsigned flags)
    {
        int quick = !(flags & HAS_OBJECT_RECHECK_PACKED);
        unsigned object_info_flags = OBJECT_INFO_SKIP_FETCH_OBJECT |
            (quick ? OBJECT_INFO_QUICK : 0);

        if (!startup_info->have_repository)
            return 0;
        return oid_object_info_extended(r, oid, NULL, object_info_flags) >= 0;
    }

    int repo_has_object_file_with_flags(struct repository *r,
                        const struct object_id *oid, int flags)
    {
        if (!startup_info->have_repository)
            return 0;
        return oid_object_info_extended(r, oid, NULL, flags) >= 0;
    }

    int repo_has_object_file(struct repository *r,
                const struct object_id *oid)
    {
        return repo_has_object_file_with_flags(r, oid, 0);
    }

Now we kown that has_object() add OBJECT_INFO_SKIP_FETCH_OBJECT to skip
fetch object.

I found that Ævar Arnfjörð Bjarmason added deref_without_lazy_fetch()
4 weeks ago[2]:
    static struct commit *deref_without_lazy_fetch(const struct object_id *oid,
                            int mark_tags_complete)
    {
        enum object_type type;
        unsigned flags = OBJECT_INFO_SKIP_FETCH_OBJECT | OBJECT_INFO_QUICK;
        return deref_without_lazy_fetch_extended(oid, mark_tags_complete,
                            &type, flags);
    }

But oi_flags is only used by oid_object_info_extended() and is missed by
lookup_commit_in_graph():
    static struct commit *deref_without_lazy_fetch_extended(const struct object_id *oid,
                                int mark_tags_complete,
                                enum object_type *type,
                                unsigned int oi_flags)
    {
        struct object_info info = { .typep = type };
        struct commit *commit;

        commit = lookup_commit_in_graph(the_repository, oid);
        if (commit)
            return commit;

        while (1) {
            if (oid_object_info_extended(the_repository, oid, &info,
                            oi_flags))

So, an appropriate fix can be that let lookup_commit_in_graph() pickup
oi_flags and pass it to oid_object_info_extended(), then the fetching
loop will be prevent by the given flag OBJECT_INFO_SKIP_FETCH_OBJECT.

1. https://github.com/git/git/blob/master/Documentation/RelNotes/2.29.0.txt
2. https://lore.kernel.org/git/2a563b5f18cc9c42cb71a9547344a5435f6bc058.1652731865.git.gitgitgadget@gmail.com/

Thanks
-Han Xin

Han Xin (2):
  commit-graph.c: add "flags" to lookup_commit_in_graph()
  fetch-pack.c: pass "oi_flags" to lookup_commit_in_graph()

 builtin/fetch.c                    |  4 ++-
 commit-graph.c                     |  5 ++--
 commit-graph.h                     |  3 +-
 fetch-pack.c                       | 10 +++----
 revision.c                         |  2 +-
 t/t5583-fetch-with-commit-graph.sh | 47 ++++++++++++++++++++++++++++++
 upload-pack.c                      |  5 ++--
 7 files changed, 64 insertions(+), 12 deletions(-)
 create mode 100644 t/t5583-fetch-with-commit-graph.sh

-- 
2.36.1


  reply	other threads:[~2022-06-16  3:39 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-06-14  7:25 An endless loop fetching issue with partial clone, alternates and commit graph Haiyng Tan
2022-06-15  2:18 ` Taylor Blau
2022-06-16  3:38   ` Han Xin [this message]
2022-06-16  3:38     ` [RFC PATCH 1/2] commit-graph.c: add "flags" to lookup_commit_in_graph() Han Xin
2022-06-16  3:38     ` [RFC PATCH 2/2] fetch-pack.c: pass "oi_flags" " Han Xin
2022-06-17 21:47     ` [RFC PATCH 0/2] Re: An endless loop fetching issue with partial clone, alternates and commit graph Jonathan Tan
2022-06-18  3:01     ` [PATCH v1] commit-graph.c: no lazy fetch in lookup_commit_in_graph() Han Xin
2022-06-20  7:07       ` Patrick Steinhardt
2022-06-20  8:53         ` [External] " 欣韩
2022-06-20  9:05           ` Patrick Steinhardt
2022-06-21 18:23       ` Jonathan Tan
2022-06-22  3:17         ` Han Xin
2022-06-24  5:27       ` [PATCH v2 0/2] " Han Xin
2022-06-24  5:27         ` [PATCH v2 1/2] test-lib.sh: add limited processes to test-lib Han Xin
2022-06-24 16:03           ` Junio C Hamano
2022-06-25  1:35             ` Han Xin
2022-06-27 12:22               ` Junio C Hamano
2022-06-24  5:27         ` [PATCH v2 2/2] commit-graph.c: no lazy fetch in lookup_commit_in_graph() Han Xin
2022-06-24 16:56           ` Junio C Hamano
2022-06-25  2:25             ` Han Xin
2022-06-25  2:31               ` Han Xin
2022-06-28  2:02         ` [PATCH v3 0/2] " Han Xin
2022-06-28  2:02           ` [PATCH v3 1/2] test-lib.sh: add limited processes to test-lib Han Xin
2022-06-28  2:02           ` [PATCH v3 2/2] commit-graph.c: no lazy fetch in lookup_commit_in_graph() Han Xin
2022-06-28  7:49             ` Ævar Arnfjörð Bjarmason
2022-06-28 17:36               ` Junio C Hamano
2022-06-30 12:21                 ` Johannes Schindelin
2022-06-30 13:43                   ` Ævar Arnfjörð Bjarmason
2022-06-30 15:40                     ` Junio C Hamano
2022-06-30 18:47                       ` Ævar Arnfjörð Bjarmason
2022-07-01 19:31                       ` Johannes Schindelin
2022-07-01 20:47                         ` Junio C Hamano
2022-06-29  2:08               ` Han Xin
2022-06-30 17:37           ` test name conflict + js/ci-github-workflow-markup regression (was: [PATCH v3 0/2] no lazy fetch in lookup_commit_in_graph()) Ævar Arnfjörð Bjarmason
2022-07-01  1:34           ` [PATCH v4 0/1] no lazy fetch in lookup_commit_in_graph() Han Xin
2022-07-01  1:34             ` [PATCH v4 1/1] commit-graph.c: " Han Xin
2022-07-09 12:23               ` Michael J Gruber
2022-07-11 15:09                 ` Jeff King
2022-07-11 20:17                   ` Junio C Hamano
2022-07-12  1:52                     ` [External] " Han Xin
2022-07-12  5:23                       ` Junio C Hamano
2022-07-12  5:32                         ` Han Xin
2022-07-12  6:37                         ` [External] " Jeff King
2022-07-12 14:19                           ` Junio C Hamano
2022-07-12  6:50             ` [PATCH v5 0/1] " Han Xin
2022-07-12  6:50               ` [PATCH v5 1/1] commit-graph.c: " Han Xin
2022-07-12  9:50                 ` Ævar Arnfjörð Bjarmason
2022-07-13  1:26                   ` Han Xin
2022-07-12  6:58               ` [PATCH v5 0/1] " Jeff King
2022-07-12  8:01             ` [PATCH v1] t5330: remove run_with_limited_processses() Han Xin

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=cover.1655350442.git.hanxin.hx@bytedance.com \
    --to=hanxin.hx@bytedance.com \
    --cc=chiyutianyi@gmail.com \
    --cc=derrickstolee@github.com \
    --cc=git@vger.kernel.org \
    --cc=haiyangtand@gmail.com \
    --cc=jonathantanmy@google.com \
    --cc=me@ttaylorr.com \
    --cc=ps@pks.im \
    /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).