From: Junio C Hamano <gitster@pobox.com>
To: Michael Haggerty <mhagger@alum.mit.edu>
Cc: "Nguyễn Thái Ngọc Duy" <pclouds@gmail.com>,
"Stefan Beller" <sbeller@google.com>, "Jeff King" <peff@peff.net>,
"Ævar Arnfjörð Bjarmason" <avarab@gmail.com>,
"David Turner" <novalis@novalis.org>,
"Brandon Williams" <bmwill@google.com>,
"Johannes Sixt" <j6t@kdbg.org>, "Øyvind Holm" <sunny@sunbase.org>,
"Linus Torvalds" <torvalds@linux-foundation.org>,
git@vger.kernel.org
Subject: Re: [PATCH 0/2] Fix a refname trimming problem in `log --bisect`
Date: Wed, 14 Jun 2017 03:09:32 -0700 [thread overview]
Message-ID: <xmqqbmpq519f.fsf@gitster.mtv.corp.google.com> (raw)
In-Reply-To: <cover.1497430232.git.mhagger@alum.mit.edu> (Michael Haggerty's message of "Wed, 14 Jun 2017 11:07:25 +0200")
Michael Haggerty <mhagger@alum.mit.edu> writes:
> The code for `git log --bisect` was calling
> `for_each_ref_in_submodule()` with prefix set to "refs/bisect/bad",
> which is the actual name of the reference that it wants. This resulted
> in the refname being trimmed completely away and the empty string
> being passed to the callback. That became impermissible after
>
> b9c8e7f2fb prefix_ref_iterator: don't trim too much, 2017-05-22
>
> , so the command was failing.
>
> Fix the problem in two orthogonal ways:
>
> 1. Add a new function, `for_each_fullref_in_submodule()`, that doesn't
> trim the refnames that it passes to callbacks, and us that instead.
> I *think* that this is a strict improvement, though I don't know
> the `git log` code well enough to be sure that it won't have bad
> side-effects.
>
> 2. Relax the "trimming too many characters" check to allow the full
> length of the refname to be trimmed away (though not more than
> that).
>
> In an ideal world the second patch shouldn't be necessary, because
> this calling pattern is questionable and it might be better that we
> learn about any other offenders. But if we'd rather be conservative
> and not break any other code that might rely on the old behavior,
> patch 2 is my suggestion for how to do it.
Thanks for a nice summary.
I agree that 2. is a nice safety to have, especially if the code
before b9c8e7f2 ("prefix_ref_iterator: don't trim too much",
2017-05-22) has been seeing the completely trimmed result (i.e. an
empty string) in the callback function.
And I agree that 1. is also a good interface to have.
next prev parent reply other threads:[~2017-06-14 10:09 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-06-14 0:06 [BUG] b9c8e7f2fb6e breaks git bisect visualize Øyvind A. Holm
2017-06-14 8:36 ` Michael Haggerty
2017-06-14 9:07 ` [PATCH 0/2] Fix a refname trimming problem in `log --bisect` Michael Haggerty
2017-06-14 9:07 ` [PATCH 1/2] for_each_bisect_ref(): don't trim refnames Michael Haggerty
2017-06-14 9:22 ` Jeff King
2017-06-14 9:32 ` Jeff King
2017-06-14 10:05 ` Junio C Hamano
2017-06-15 22:26 ` Junio C Hamano
2017-06-14 9:07 ` [PATCH 2/2] prefix_ref_iterator_advance(): relax the check of trim length Michael Haggerty
2017-06-14 9:24 ` [PATCH 0/2] Fix a refname trimming problem in `log --bisect` Jeff King
2017-06-14 10:09 ` Junio C Hamano [this message]
2017-06-14 9:18 ` [BUG] b9c8e7f2fb6e breaks git bisect visualize Jeff King
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=xmqqbmpq519f.fsf@gitster.mtv.corp.google.com \
--to=gitster@pobox.com \
--cc=avarab@gmail.com \
--cc=bmwill@google.com \
--cc=git@vger.kernel.org \
--cc=j6t@kdbg.org \
--cc=mhagger@alum.mit.edu \
--cc=novalis@novalis.org \
--cc=pclouds@gmail.com \
--cc=peff@peff.net \
--cc=sbeller@google.com \
--cc=sunny@sunbase.org \
--cc=torvalds@linux-foundation.org \
/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.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.