git@vger.kernel.org list mirror (unofficial, one of many)
 help / color / mirror / code / Atom feed
From: Denton Liu <liu.denton@gmail.com>
To: Junio C Hamano <gitster@pobox.com>
Cc: Git Mailing List <git@vger.kernel.org>
Subject: Re: [PATCH v3 06/10] diff-lib: define diff_get_merge_base()
Date: Fri, 18 Sep 2020 03:34:00 -0700	[thread overview]
Message-ID: <20200918103400.GA1874074@generichostname> (raw)
In-Reply-To: <xmqqd02ks4qk.fsf@gitster.c.googlers.com>

Hi Junio,

On Thu, Sep 17, 2020 at 10:16:51AM -0700, Junio C Hamano wrote:
> Denton Liu <liu.denton@gmail.com> writes:
> 
> > +void diff_get_merge_base(const struct rev_info *revs, struct object_id *mb)
> > +{
> > +	int i;
> > +	struct commit *mb_child[2] = {0};
> > +	struct commit_list *merge_bases;
> > +
> > +	for (i = 0; i < revs->pending.nr; i++) {
> > +		struct object *obj = revs->pending.objects[i].item;
> > +		if (obj->flags)
> > +			die(_("--merge-base does not work with ranges"));
> > +		if (obj->type != OBJ_COMMIT)
> > +			die(_("--merge-base only works with commits"));
> > +	}
> 
> This is the first use of die() in this file, that is designed to
> keep a set of reusable library functions so that the caller(s) can
> do their own die().  They may want to become

Although this is the first instance of die(), run_diff_index() has an
exit(128), which is basically a die() in disguise.

> 	return error(_(...));
> 
> The same comment applies to all die()s the patch adds.

I applied this change but then each callsite of diff_get_merge_base()
became something like

	if (diff_get_merge_base(revs, &oid))
		exit(128);

so I do agree with the spirit of the change but in reality, it just
creates more busywork for the callers.

> > +	/*
> > +	 * This check must go after the for loop above because A...B
> > +	 * ranges produce three pending commits, resulting in a
> > +	 * misleading error message.
> > +	 */
> 
> Should "git diff --merge-base A...B" be forbidden, or does it just
> ask the same thing twice and is not a die-worthy offence?

I think that it should be die-worthy because it's a logic error for a
user to do this. I can't think of any situation where it wouldn't be
more desirable error early to correct a user's thinking. Plus, we're
trying to move away from the `...` notation anyway ;)

> > +	for (i = 0; i < revs->pending.nr; i++)
> > +		mb_child[i] = lookup_commit_reference(the_repository, &revs->pending.objects[i].item->oid);
> > +	if (revs->pending.nr < ARRAY_SIZE(mb_child)) {
> > +		struct object_id oid;
> > +
> > +		if (revs->pending.nr != 1)
> > +			BUG("unexpected revs->pending.nr: %d", revs->pending.nr);
> 
> This is an obviously impossible condition as we will not take more
> than 2.

We also want to ensure that revs->pending.nr isn't 0 here. That being
said, I can explicitly check earlier in the function that the number of
pending is 1 or 2 so that it's more clearly written.

Thanks,
Denton

  reply	other threads:[~2020-09-18 10:34 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-05 19:08 [PATCH 0/4] builtin/diff: learn --merge-base Denton Liu
2020-09-05 19:08 ` [PATCH 1/4] t4068: remove unnecessary >tmp Denton Liu
2020-09-05 19:08 ` [PATCH 2/4] git-diff.txt: backtick quote command text Denton Liu
2020-09-05 19:08 ` [PATCH 3/4] builtin/diff: parse --no-index using parse_options() Denton Liu
2020-09-05 19:08 ` [PATCH 4/4] builtin/diff: learn --merge-base Denton Liu
2020-09-06 21:58   ` Junio C Hamano
2020-09-10  7:32 ` [PATCH v2 0/4] " Denton Liu
2020-09-10  7:32   ` [PATCH v2 1/4] t4068: remove unnecessary >tmp Denton Liu
2020-09-10  7:32   ` [PATCH v2 2/4] git-diff.txt: backtick quote command text Denton Liu
2020-09-10  7:32   ` [PATCH v2 3/4] builtin/diff: parse --no-index using parse_options() Denton Liu
2020-09-10 18:35     ` Junio C Hamano
2020-09-13  8:31       ` Denton Liu
2020-09-13 21:45         ` Junio C Hamano
2020-09-10  7:32   ` [PATCH v2 4/4] builtin/diff: learn --merge-base Denton Liu
2020-09-17  7:44   ` [PATCH v3 00/10] " Denton Liu
2020-09-17  7:44     ` [PATCH v3 01/10] t4068: remove unnecessary >tmp Denton Liu
2020-09-17  7:44     ` [PATCH v3 02/10] git-diff-index.txt: make --cached description a proper sentence Denton Liu
2020-09-17  7:44     ` [PATCH v3 03/10] git-diff.txt: backtick quote command text Denton Liu
2020-09-17  7:44     ` [PATCH v3 04/10] contrib/completion: extract common diff/difftool options Denton Liu
2020-09-17  7:44     ` [PATCH v3 05/10] diff-lib: accept option flags in run_diff_index() Denton Liu
2020-09-17 17:00       ` Junio C Hamano
2020-09-17  7:44     ` [PATCH v3 06/10] diff-lib: define diff_get_merge_base() Denton Liu
2020-09-17 17:16       ` Junio C Hamano
2020-09-18 10:34         ` Denton Liu [this message]
2020-09-19  0:33           ` Junio C Hamano
2020-09-17  7:44     ` [PATCH v3 07/10] t4068: add --merge-base tests Denton Liu
2020-09-17  7:44     ` [PATCH v3 08/10] builtin/diff-index: learn --merge-base Denton Liu
2020-09-17 17:28       ` Junio C Hamano
2020-09-17 18:13         ` Jeff King
2020-09-18  5:11           ` Junio C Hamano
2020-09-18 18:12             ` Jeff King
2020-09-17  7:44     ` [PATCH v3 09/10] builtin/diff-tree: " Denton Liu
2020-09-17 18:23       ` Junio C Hamano
2020-09-18 10:48         ` Denton Liu
2020-09-18 16:52           ` Junio C Hamano
2020-09-20 11:01             ` Denton Liu
2020-09-21 16:05               ` Junio C Hamano
2020-09-21 17:27                 ` Denton Liu
2020-09-21 21:09                   ` Junio C Hamano
2020-09-21 21:19                     ` Junio C Hamano
2020-09-21 21:54                     ` Denton Liu
2020-09-21 22:18                       ` Junio C Hamano
2020-09-23  9:47                         ` Denton Liu
2020-09-25 21:02                           ` Junio C Hamano
2020-09-26  1:52                             ` Denton Liu
2020-09-17  7:44     ` [PATCH v3 10/10] contrib/completion: complete `git diff --merge-base` Denton Liu
2020-09-20 11:22     ` [PATCH v4 00/10] builtin/diff: learn --merge-base Denton Liu
2020-09-20 11:22       ` [PATCH v4 01/10] t4068: remove unnecessary >tmp Denton Liu
2020-09-20 11:22       ` [PATCH v4 02/10] git-diff-index.txt: make --cached description a proper sentence Denton Liu
2020-09-20 11:22       ` [PATCH v4 03/10] git-diff.txt: backtick quote command text Denton Liu
2020-09-20 11:22       ` [PATCH v4 04/10] contrib/completion: extract common diff/difftool options Denton Liu
2020-09-20 11:22       ` [PATCH v4 05/10] diff-lib: accept option flags in run_diff_index() Denton Liu
2020-09-20 11:22       ` [PATCH v4 06/10] diff-lib: define diff_get_merge_base() Denton Liu
2020-09-20 11:22       ` [PATCH v4 07/10] t4068: add --merge-base tests Denton Liu
2020-09-20 11:22       ` [PATCH v4 08/10] builtin/diff-index: learn --merge-base Denton Liu
2020-09-29 19:53         ` Martin Ågren
2020-09-20 11:22       ` [PATCH v4 09/10] builtin/diff-tree: " Denton Liu
2020-09-20 11:22       ` [PATCH v4 10/10] contrib/completion: complete `git diff --merge-base` Denton Liu

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=20200918103400.GA1874074@generichostname \
    --to=liu.denton@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --subject='Re: [PATCH v3 06/10] diff-lib: define diff_get_merge_base()' \
    /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).