git@vger.kernel.org list mirror (unofficial, one of many)
 help / color / mirror / code / Atom feed
From: Kaartic Sivaraam <kaartic.sivaraam@gmail.com>
To: Junio C Hamano <gitster@pobox.com>
Cc: Shourya Shukla <shouryashukla.oo@gmail.com>,
	Johannes.Schindelin@gmx.de, chriscool@tuxfamily.org,
	christian.couder@gmail.com, git@vger.kernel.org,
	liu.denton@gmail.com, pc44800@gmail.com, stefanbeller@gmail.com
Subject: Re: [PATCH v3 4/4] submodule: port submodule subcommand 'summary' from shell to C
Date: Mon, 24 Aug 2020 01:33:17 +0530	[thread overview]
Message-ID: <5b6ed82f3ab58a194bf51c0e2905214f64246ad8.camel@gmail.com> (raw)
In-Reply-To: <xmqqa6yn93ll.fsf@gitster.c.googlers.com>

On Fri, 2020-08-21 at 12:54 -0700, Junio C Hamano wrote:
> Kaartic Sivaraam <kaartic.sivaraam@gmail.com> writes:
> 
> > Here's the error message with context of the trash directory of that
> > test:
> > 
> > -- 8< --
> > $ cd t
> > $ ./t7421-submodule-summary-add.sh  -d
> > $ cd trash\ directory.t7421-submodule-summary-add/
> > 
> > $ git submodule summary HEAD^^
> > fatal: exec 'rev-parse': cd to 'my-subm' failed: No such file or directory
> > * my-subm 35b40f1...0000000:
> > 
> > * subm 0000000...dbd5fc8 (2):
> >   > add file2
> > 
> > -- >8 --
> > 
> > That 'fatal' is a consequence of spawning a process in
> > `verify_submodule_committish` of `builtin/submodule--helper.c` even for
> > non-existent submodules.
> 
> Oh, so doing something that would cause the error message to be
> emitted itself is a bug.
> 

Exactly.

> > I don't think that 'fatal: ' message is giving
> > any useful information here. The fact that submodule 'my-subm' doesn't
> > exist can easily be inferred just by looking at the destination mode
> > (0000000). If anything that 'fatal' message is just confusing and
> > unnecessary, IMO.
> 
> Yes, I 100% agree.
> 
> > So, we could easily suppress it by doing something like this (while
> > also fixing the test):
> 
> Yup.  That is a very good idea.  
> 
> Or the caller of verify_submodule_committish() should refrain from
> calling it for the path?  After all, it is checking sm_path is a
> path to where a submodule should be before calling the function
> (instead of calling it for every random path), iow its criteria to
> make the call currently may be "the path in the index says it is a
> submodule", but it should easily be updated to "the path in the
> index says it is a submodule, and the submodule actually is
> populated", right?
> 

Ah, this reminds me of the initial version of the patch which did
exactly that. Quoting it here for reference:

+	strbuf_addstr(&sm_git_dir_sb, p->sm_path);
+	if (is_nonbare_repository_dir(&sm_git_dir_sb))
+		is_sm_git_dir = 1;
+
+	if (is_sm_git_dir && S_ISGITLINK(p->mod_src))
+		missing_src = verify_submodule_object_name(p->sm_path,
+							   oid_to_hex(&p->oid_src));
+
+	if (is_sm_git_dir && S_ISGITLINK(p->mod_dst))
+		missing_dst = verify_submodule_object_name(p->sm_path,
+							   oid_to_hex(&p->oid_dst));
+

Note: `verify_submodule_object_name` is now renamed to
`verify_submodule_committish`.

That does sound like a sane approach to me. There's not much point in
invoking `rev-parse` in a non-populated (a.k.a. de-initialized) or non-
existent submodule but we removed that check as we thought it was
unnecessary redundant because `capture_command` would fail anyway.
Looks like we failed to notice the additional `fatal` message fallout
then.

Also, I think it would be better to something like the following in
t7421 to ensure that `fatal` doesn't sneak up accidentally in the
future:

-- 8< --
diff --git t/t7421-submodule-summary-add.sh t/t7421-submodule-summary-add.sh
index 59a9b00467..b070f13714 100755
--- t/t7421-submodule-summary-add.sh
+++ t/t7421-submodule-summary-add.sh
@@ -58,7 +58,7 @@ test_expect_success 'submodule summary output for submodules with changed paths'
        git commit -m "change submodule path" &&
        rev=$(git -C sm rev-parse --short HEAD^) &&
        git submodule summary HEAD^^ -- my-subm >actual 2>err &&
-       grep "fatal:.*my-subm" err &&
+       test_must_be_empty err &&
        cat >expected <<-EOF &&
        * my-subm ${rev}...0000000:
 
-- >8 --

-- 
Sivaraam



  reply	other threads:[~2020-08-23 20:03 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-06 16:40 [GSoC][PATCH v2 0/5] submodule: port " Shourya Shukla
2020-08-06 16:40 ` [PATCH v2 1/5] submodule: expose the '--for-status' option of summary Shourya Shukla
2020-08-08 14:40   ` Kaartic Sivaraam
2020-08-08 20:25     ` Christian Couder
2020-08-08 23:26       ` Junio C Hamano
2020-08-06 16:40 ` [PATCH v2 2/5] submodule: remove extra line feeds between callback struct and macro Shourya Shukla
2020-08-06 16:41 ` [PATCH v2 3/5] submodule: rename helper functions to avoid ambiguity Shourya Shukla
2020-08-06 16:41 ` [PATCH v2 4/5] t7421: introduce a test script for verifying 'summary' output Shourya Shukla
2020-08-06 16:41 ` [PATCH v2 5/5] submodule: port submodule subcommand 'summary' from shell to C Shourya Shukla
2020-08-06 22:45   ` Junio C Hamano
2020-08-07 16:31     ` Shourya Shukla
2020-08-07 17:15       ` Junio C Hamano
2020-08-12 19:44 ` [GSoC][PATCH v3 0/4] submodule: port " Shourya Shukla
2020-08-12 19:44   ` [PATCH v3 1/4] submodule: remove extra line feeds between callback struct and macro Shourya Shukla
2020-08-12 19:44   ` [PATCH v3 2/4] submodule: rename helper functions to avoid ambiguity Shourya Shukla
2020-08-12 19:44   ` [PATCH v3 3/4] t7421: introduce a test script for verifying 'summary' output Shourya Shukla
2020-08-12 19:44   ` [PATCH v3 4/4] submodule: port submodule subcommand 'summary' from shell to C Shourya Shukla
2020-08-18  2:08     ` Jeff King
2020-08-21  5:22       ` Shourya Shukla
2020-08-21 15:17     ` Johannes Schindelin
2020-08-21 16:35       ` Junio C Hamano
2020-08-21 17:17         ` Shourya Shukla
2020-08-21 18:09           ` Junio C Hamano
2020-08-21 18:54             ` Kaartic Sivaraam
2020-08-21 19:54               ` Junio C Hamano
2020-08-23 20:03                 ` Kaartic Sivaraam [this message]
2020-08-23 20:12                   ` Kaartic Sivaraam
2020-08-24  7:26                   ` Shourya Shukla
2020-08-24  8:46                     ` Shourya Shukla
2020-08-24 11:08                       ` Kaartic Sivaraam
2020-08-24 17:50                         ` Shourya Shukla
2020-08-24 17:54     ` Junio C Hamano

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=5b6ed82f3ab58a194bf51c0e2905214f64246ad8.camel@gmail.com \
    --to=kaartic.sivaraam@gmail.com \
    --cc=Johannes.Schindelin@gmx.de \
    --cc=chriscool@tuxfamily.org \
    --cc=christian.couder@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=liu.denton@gmail.com \
    --cc=pc44800@gmail.com \
    --cc=shouryashukla.oo@gmail.com \
    --cc=stefanbeller@gmail.com \
    --subject='Re: [PATCH v3 4/4] submodule: port submodule subcommand '\''summary'\'' from shell to C' \
    /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

Code repositories for project(s) associated with this 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).