From: Junio C Hamano <gitster@pobox.com>
To: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Cc: git@vger.kernel.org, Robbie Iannucci <iannucci@google.com>
Subject: Re: [PATCH 3/3] lockfile: LOCK_REPORT_ON_ERROR
Date: Thu, 08 Dec 2016 10:22:06 -0800 [thread overview]
Message-ID: <xmqqinquth69.fsf@gitster.mtv.corp.google.com> (raw)
In-Reply-To: <alpine.DEB.2.20.1612081252490.23160@virtualbox> (Johannes Schindelin's message of "Thu, 8 Dec 2016 12:53:39 +0100 (CET)")
Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:
> Sorry for the breakage.
Apologies from me, too. Once a topic is merged, the credit still
remains with the contributor, but the blame is shared by the project
as a whole, with those who missed breakages during their reviews,
and those who didn't review or test and let breakages pass.
> When libifying the code, I tried to be careful to retain the error
> messages when not dying,...
Quite honestly, I do not think either of us cared about preserving
the exact error message the end-user was getting from each failure
sites that the series changed a call with die-on-error=1 to a call
with die-on-error=0 that is followed by a negative return while
reviewing this series. As I wrote in the proposed log message for
3/3, this one was noticed as a end-user breaking change because it
was the only one that has become totally silent. For example, this
bit from sequencer.c::write_message() we can see in the output from
"git show --first-parent 2a4062a4a8" does not preserve the message
at all:
diff --git a/sequencer.c b/sequencer.c
index 3804fa931d..eec8a60d6b 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -180,17 +180,20 @@
...
-static void write_message(struct strbuf *msgbuf, const char *filename)
+static int write_message(struct strbuf *msgbuf, const char *filename)
{
static struct lock_file msg_file;
- int msg_fd = hold_lock_file_for_update(&msg_file, filename,
- LOCK_DIE_ON_ERROR);
+ int msg_fd = hold_lock_file_for_update(&msg_file, filename, 0);
+ if (msg_fd < 0)
+ return error_errno(_("Could not lock '%s'"), filename);
And I do not think it is necessarily bad that the error message
changed with this conversion. In other words, I do not think it
should have been the goal to preserve the exact error message.
hold_lock*() can afford to give a detailed message that strongly
sounds as being the final decision when called with die-on-error=1
because it knows it is dying. However, the message from the updated
write_message(), "could not lock", cannot be final---the caller may
want to add something else after it to describe what failed in a
larger picture.
prev parent reply other threads:[~2016-12-08 18:22 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-12-03 1:44 [BUG] Index.lock error message regression in git 2.11.0 Robbie Iannucci
2016-12-03 1:52 ` Robbie Iannucci
2016-12-06 21:56 ` Junio C Hamano
2016-12-06 22:15 ` Re* " Junio C Hamano
2016-12-07 18:21 ` Junio C Hamano
2016-12-07 19:41 ` [PATCH 0/3] Do not be totally silent upon lock error Junio C Hamano
2016-12-07 19:41 ` [PATCH 1/3] wt-status: implement opportunisitc index update correctly Junio C Hamano
2016-12-07 20:48 ` Stefan Beller
2016-12-07 20:53 ` Junio C Hamano
2016-12-07 20:56 ` Stefan Beller
2016-12-07 21:08 ` Junio C Hamano
2016-12-07 22:06 ` Stefan Beller
2016-12-08 10:18 ` Paul Tan
2016-12-08 18:01 ` Junio C Hamano
2016-12-07 19:41 ` [PATCH 2/3] hold_locked_index(): align error handling with hold_lockfile_for_update() Junio C Hamano
2016-12-07 19:41 ` [PATCH 3/3] lockfile: LOCK_REPORT_ON_ERROR Junio C Hamano
2016-12-08 11:53 ` Johannes Schindelin
2016-12-08 18:10 ` Robbie Iannucci
2016-12-08 18:22 ` Junio C Hamano [this message]
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=xmqqinquth69.fsf@gitster.mtv.corp.google.com \
--to=gitster@pobox.com \
--cc=Johannes.Schindelin@gmx.de \
--cc=git@vger.kernel.org \
--cc=iannucci@google.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).