From: Phillip Wood <phillip.wood123@gmail.com>
To: Denton Liu <liu.denton@gmail.com>, phillip.wood@dunelm.org.uk
Cc: Git Mailing List <git@vger.kernel.org>,
Alban Gruin <alban.gruin@gmail.com>,
Johannes Schindelin <johannes.schindelin@gmx.de>,
Junio C Hamano <gitster@pobox.com>
Subject: Re: [PATCH v3 17/19] merge: teach --autostash option
Date: Fri, 3 Apr 2020 14:34:30 +0100 [thread overview]
Message-ID: <194df5e2-4e5d-3150-d8a4-bb55bdef75be@gmail.com> (raw)
In-Reply-To: <20200403103126.GA3163329@generichostname>
Hi Denton
On 03/04/2020 11:31, Denton Liu wrote:
> Hi Phillip,
>
> Thanks for your detailed review on the series. I've updated everything
> according to your comments except for the parts below:
That's great, I've added a few more comments below
> On Thu, Apr 02, 2020 at 04:24:54PM +0100, Phillip Wood wrote:
>>> diff --git a/branch.c b/branch.c
>>> index 579494738a..bf2536c70d 100644
>>> --- a/branch.c
>>> +++ b/branch.c
>>> @@ -344,6 +344,7 @@ void remove_merge_branch_state(struct repository *r)
>>> unlink(git_path_merge_rr(r));
>>> unlink(git_path_merge_msg(r));
>>> unlink(git_path_merge_mode(r));
>>> + apply_autostash(git_path_merge_autostash(r));
>>> }
>>> void remove_branch_state(struct repository *r, int verbose)
>>> diff --git a/builtin/commit.c b/builtin/commit.c
>>> index 7ba33a3bec..c11894423a 100644
>>> --- a/builtin/commit.c
>>> +++ b/builtin/commit.c
>>> @@ -1687,6 +1687,8 @@ int cmd_commit(int argc, const char **argv, const char *prefix)
>>> unlink(git_path_merge_mode(the_repository));
>>> unlink(git_path_squash_msg(the_repository));
>>> + apply_autostash(git_path_merge_autostash(the_repository));
>>> +
>>
>> It's hard to tell from the limited context but do we want to run
>> commit_index_files() before applying the autostash?
>
> I don't think it really matters which order we run it in. When we run
> apply_autostash(), we only ever touch the working tree, not the index so
> it doesn't matter if it's run before or after. I'd prefer to keep it
> here because if we ever refactor this to use
> remove_merge_branch_state(), the person working on this will be able to
> perform the refactor more easily without having to worry about implicit
> ordering dependencies.
Thinking a bit more about this we definitely want to commit the index
files before applying the stash as that is done in a separate process so
we want our index written to disk first.
'git stash apply' does touch the index while it's merging and then tries
to reset it to the pre-merge state if the merge has no conflicts. If the
user runs
git add new-file
git merge --autostash branch
and new-file is not in branch then 'git stash apply' will add new-file
to the index
>> I wonder if this should
>> be using remove_merge_branch_state() instead of duplicating it as well.
>
> We can _almost_ use remove_branch_state() even! Unfortunately,
> remove_merge_branch_state() calls `unlink(git_path_merge_rr(r))` which
> we cannot do since repo_rerere() relies on it later. Perhaps we can
> can move repo_rerere() earlier?
I think we should move apply_autostash() down so that repo_rerere() is
called with the index that we've committed before we apply the stash
which might add conflicts or new files. We should probably run the
post-commit and post-rewrite hooks before we apply the autostash as well
to avoid changing the existing behaviour.
If we aim for something like
commit_index_files()
repo_rerere()
stash_oid = read_autostash_oid()
cleanup_branch_state()
run_post_commit_hook()
run_post_rewrite_hook()
print_commit_summary()
apply_autostash(stash_oid)
Then we keep the existing behaviour for rerere and the hooks, and the
commit summary is printed before all the status junk that
apply_autostash() spews out
We probably need to check the where 'git merge' applies the autostash as
well to make sure it is writing the index to disk, calling the
post-merge hook and printing any messages before applying the stash
Best Wishes
Phillip
>>> diff --git a/builtin/reset.c b/builtin/reset.c
>>> index 18228c312e..038c8532eb 100644
>>> --- a/builtin/reset.c
>>> +++ b/builtin/reset.c
>>> @@ -25,6 +25,7 @@
>>> #include "cache-tree.h"
>>> #include "submodule.h"
>>> #include "submodule-config.h"
>>> +#include "sequencer.h"
>>> #define REFRESH_INDEX_DELAY_WARNING_IN_MS (2 * 1000)
>>> @@ -437,8 +438,12 @@ int cmd_reset(int argc, const char **argv, const char *prefix)
>>> if (reset_type == HARD && !update_ref_status && !quiet)
>>> print_new_head_line(lookup_commit_reference(the_repository, &oid));
>>> }
>>> - if (!pathspec.nr)
>>> + if (!pathspec.nr) {
>>> + if (reset_type == HARD)
>>> + save_autostash(git_path_merge_autostash(the_repository));
>>> +
>>> remove_branch_state(the_repository, 0);
>>
>> This removes the autostash file for all reset types but we only keep the
>> stash in the case of 'reset --hard' which is confusing.
>
> I was worried that this change would be controversial... The rationale
> behind this change was that with `reset --hard`, we want to leave a
> clean working tree behind so we save it into the stash reflog. In all
> other cases, remove_branch_state() will apply the saved stash entry
> which should be fine since users don't expect a clean worktree with the
> other reset types.
>
> I considered saving the autostash in all cases of reset but
> `git merge --abort` invokes `git reset --merge` behind the scenes so
> we'd have to consider that. Perhaps we can make all resets save the
> stash entry and, in the case of `merge --abort`, we can add some extra
> logic to subvert this so that the stash entry is applied?
>
> I'm not sure about what the most intuitive behaviour is. I thought that
> this implementation would be the best but I'm not entirely sure. I'd
> appreciate some more discussion on this.
>
> Thanks,
>
> Denton
>
>> Best Wishes
>>
>> Phillip
>>
>>> + }
>>> return update_ref_status;
>>> }
next prev parent reply other threads:[~2020-04-03 13:34 UTC|newest]
Thread overview: 152+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-16 17:26 [RFC PATCH 0/7] merge: learn --autostash Denton Liu
2019-10-16 17:26 ` [RFC PATCH 1/7] Makefile: alphabetically sort += lists Denton Liu
2019-10-17 18:07 ` Junio C Hamano
2019-10-21 18:44 ` Johannes Schindelin
2019-10-21 18:53 ` Denton Liu
2019-10-22 23:33 ` Johannes Schindelin
2019-10-21 19:49 ` Junio C Hamano
2019-10-21 19:54 ` Jeff King
2019-10-22 11:34 ` Junio C Hamano
2019-10-16 17:26 ` [RFC PATCH 2/7] autostash: extract read_one() from rebase Denton Liu
2019-10-18 9:04 ` Phillip Wood
2019-10-21 18:46 ` Johannes Schindelin
2019-10-16 17:26 ` [RFC PATCH 3/7] autostash: extract apply_autostash() " Denton Liu
2019-10-16 17:26 ` [RFC PATCH 4/7] autostash: extract reset_head() " Denton Liu
2019-10-18 9:37 ` Phillip Wood
2019-10-21 18:56 ` Johannes Schindelin
2019-10-16 17:26 ` [RFC PATCH 5/7] autostash: extract perform_autostash() " Denton Liu
2019-10-21 19:00 ` Johannes Schindelin
2019-10-16 17:26 ` [RFC PATCH 6/7] autostash.c: undefine USE_THE_INDEX_COMPATIBILITY_MACROS Denton Liu
2019-10-18 9:41 ` Phillip Wood
2019-10-16 17:26 ` [RFC PATCH 7/7] merge: teach --autostash option Denton Liu
2019-10-18 9:46 ` Phillip Wood
2019-12-24 9:58 ` Denton Liu
2019-10-21 19:10 ` Johannes Schindelin
2019-12-24 10:05 ` Denton Liu
2019-12-24 10:12 ` Denton Liu
2019-12-24 11:04 ` [PATCH v2 00/17] merge: learn --autostash Denton Liu
2019-12-24 11:04 ` [PATCH v2 01/17] Makefile: alphabetically sort += lists Denton Liu
2019-12-30 21:38 ` Junio C Hamano
2019-12-24 11:04 ` [PATCH v2 02/17] t7600: use test_write_lines() Denton Liu
2019-12-24 11:05 ` [PATCH v2 03/17] sequencer: use file strbuf for read_oneliner() Denton Liu
2019-12-24 11:05 ` [PATCH v2 04/17] sequencer: configurably warn on non-existent files Denton Liu
2019-12-26 20:39 ` Junio C Hamano
2019-12-24 11:05 ` [PATCH v2 05/17] sequencer: make read_oneliner() extern Denton Liu
2019-12-24 11:05 ` [PATCH v2 06/17] rebase: use read_oneliner() Denton Liu
2019-12-24 11:05 ` [PATCH v2 07/17] sequencer: make apply_rebase() accept a path Denton Liu
2019-12-24 11:05 ` [PATCH v2 08/17] rebase: use apply_autostash() from sequencer.c Denton Liu
2019-12-24 11:05 ` [PATCH v2 09/17] rebase: generify reset_head() Denton Liu
2019-12-24 11:05 ` [PATCH v2 10/17] reset: extract reset_head() from rebase Denton Liu
2019-12-24 11:05 ` [PATCH v2 11/17] rebase: extract create_autostash() Denton Liu
2019-12-24 11:05 ` [PATCH v2 12/17] rebase: generify create_autostash() Denton Liu
2019-12-24 11:05 ` [PATCH v2 13/17] sequencer: extract perform_autostash() from rebase Denton Liu
2019-12-24 11:05 ` [PATCH v2 14/17] sequencer: unlink autostash in apply_autostash() Denton Liu
2019-12-24 11:05 ` [PATCH v2 15/17] merge: teach --autostash option Denton Liu
2019-12-24 11:05 ` [PATCH v2 16/17] t5520: make test_pull_autostash() accept expect_parent_num Denton Liu
2019-12-24 11:05 ` [PATCH v2 17/17] pull: pass --autostash to merge Denton Liu
2019-12-30 21:49 ` [PATCH v2 00/17] merge: learn --autostash Junio C Hamano
2019-12-31 10:34 ` Phillip Wood
2020-01-08 6:08 ` Denton Liu
2020-01-10 14:44 ` Phillip Wood
2020-01-15 16:20 ` Denton Liu
2020-01-01 7:48 ` Denton Liu
2020-03-21 9:21 ` [PATCH v3 00/19] " Denton Liu
2020-03-21 9:21 ` [PATCH v3 01/19] Makefile: ASCII-sort += lists Denton Liu
2020-03-21 9:21 ` [PATCH v3 02/19] t7600: use test_write_lines() Denton Liu
2020-03-21 9:21 ` [PATCH v3 03/19] sequencer: use file strbuf for read_oneliner() Denton Liu
2020-04-02 13:34 ` Phillip Wood
2020-03-21 9:21 ` [PATCH v3 04/19] sequencer: make read_oneliner() accept flags Denton Liu
2020-03-21 9:21 ` [PATCH v3 05/19] sequencer: configurably warn on non-existent files Denton Liu
2020-04-02 13:39 ` Phillip Wood
2020-03-21 9:21 ` [PATCH v3 06/19] sequencer: make read_oneliner() extern Denton Liu
2020-03-21 9:21 ` [PATCH v3 07/19] rebase: use read_oneliner() Denton Liu
2020-04-02 13:41 ` Phillip Wood
2020-03-21 9:21 ` [PATCH v3 08/19] sequencer: make apply_rebase() accept a path Denton Liu
2020-03-21 9:21 ` [PATCH v3 09/19] rebase: use apply_autostash() from sequencer.c Denton Liu
2020-04-02 14:59 ` Phillip Wood
2020-03-21 9:21 ` [PATCH v3 10/19] rebase: generify reset_head() Denton Liu
2020-03-21 9:21 ` [PATCH v3 11/19] reset: extract reset_head() from rebase Denton Liu
2020-04-02 15:04 ` Phillip Wood
2020-03-21 9:21 ` [PATCH v3 12/19] rebase: extract create_autostash() Denton Liu
2020-03-21 9:21 ` [PATCH v3 13/19] rebase: generify create_autostash() Denton Liu
2020-03-21 9:21 ` [PATCH v3 14/19] sequencer: extract perform_autostash() from rebase Denton Liu
2020-03-21 9:21 ` [PATCH v3 15/19] sequencer: unlink autostash in apply_autostash() Denton Liu
2020-03-21 9:21 ` [PATCH v3 16/19] sequencer: implement save_autostash() Denton Liu
2020-04-02 15:10 ` Phillip Wood
2020-03-21 9:21 ` [PATCH v3 17/19] merge: teach --autostash option Denton Liu
2020-04-02 15:24 ` Phillip Wood
2020-04-03 10:31 ` Denton Liu
2020-04-03 10:56 ` Denton Liu
2020-04-03 13:09 ` Phillip Wood
2020-04-03 21:14 ` Denton Liu
2020-04-03 13:34 ` Phillip Wood [this message]
2020-04-03 22:25 ` Denton Liu
2020-03-21 9:21 ` [PATCH v3 18/19] t5520: make test_pull_autostash() accept expect_parent_num Denton Liu
2020-03-21 9:21 ` [PATCH v3 19/19] pull: pass --autostash to merge Denton Liu
2020-04-02 16:07 ` Phillip Wood
2020-04-04 1:11 ` [PATCH v4 00/23] merge: learn --autostash Denton Liu
2020-04-04 1:11 ` [PATCH v4 01/23] Makefile: ASCII-sort += lists Denton Liu
2020-04-04 1:11 ` [PATCH v4 02/23] t7600: use test_write_lines() Denton Liu
2020-04-04 1:11 ` [PATCH v4 03/23] sequencer: stop leaking buf Denton Liu
2020-04-05 21:33 ` Junio C Hamano
2020-04-05 21:37 ` Junio C Hamano
2020-04-05 23:42 ` Denton Liu
2020-04-04 1:11 ` [PATCH v4 04/23] sequencer: reuse strbuf_trim() in read_oneliner() Denton Liu
2020-04-05 21:46 ` Junio C Hamano
2020-04-06 1:39 ` Denton Liu
2020-04-06 14:03 ` Phillip Wood
2020-04-06 14:42 ` Phillip Wood
2020-04-07 13:56 ` Denton Liu
2020-04-04 1:11 ` [PATCH v4 05/23] sequencer: make file exists check more efficient Denton Liu
2020-04-04 1:11 ` [PATCH v4 06/23] sequencer: make read_oneliner() accept flags Denton Liu
2020-04-04 1:11 ` [PATCH v4 07/23] sequencer: configurably warn on non-existent files Denton Liu
2020-04-06 14:45 ` Phillip Wood
2020-04-04 1:11 ` [PATCH v4 08/23] sequencer: make read_oneliner() extern Denton Liu
2020-04-04 1:11 ` [PATCH v4 09/23] rebase: use read_oneliner() Denton Liu
2020-04-04 1:11 ` [PATCH v4 10/23] sequencer: make apply_rebase() accept a path Denton Liu
2020-04-06 15:07 ` Phillip Wood
2020-04-04 1:11 ` [PATCH v4 11/23] sequencer: rename stash_sha1 to stash_oid Denton Liu
2020-04-04 1:11 ` [PATCH v4 12/23] rebase: use apply_autostash() from sequencer.c Denton Liu
2020-04-04 1:11 ` [PATCH v4 13/23] rebase: generify reset_head() Denton Liu
2020-04-04 1:11 ` [PATCH v4 14/23] reset: extract reset_head() from rebase Denton Liu
2020-04-04 1:11 ` [PATCH v4 15/23] rebase: extract create_autostash() Denton Liu
2020-04-04 1:11 ` [PATCH v4 16/23] rebase: generify create_autostash() Denton Liu
2020-04-04 1:11 ` [PATCH v4 17/23] sequencer: extract perform_autostash() from rebase Denton Liu
2020-04-04 1:11 ` [PATCH v4 18/23] sequencer: unlink autostash in apply_autostash() Denton Liu
2020-04-04 1:11 ` [PATCH v4 19/23] sequencer: implement save_autostash() Denton Liu
2020-04-06 15:15 ` Phillip Wood
2020-04-04 1:11 ` [PATCH v4 20/23] sequencer: implement apply_autostash_oid() Denton Liu
2020-04-04 1:11 ` [PATCH v4 21/23] merge: teach --autostash option Denton Liu
2020-04-06 15:20 ` Phillip Wood
2020-04-07 13:09 ` Denton Liu
2020-04-07 15:06 ` Phillip Wood
2020-04-04 1:11 ` [PATCH v4 22/23] t5520: make test_pull_autostash() accept expect_parent_num Denton Liu
2020-04-04 1:11 ` [PATCH v4 23/23] pull: pass --autostash to merge Denton Liu
2020-04-07 14:27 ` [PATCH v5 00/22] merge: learn --autostash Denton Liu
2020-04-07 14:27 ` [PATCH v5 01/22] Makefile: ASCII-sort += lists Denton Liu
2020-04-07 14:27 ` [PATCH v5 02/22] t7600: use test_write_lines() Denton Liu
2020-04-07 14:27 ` [PATCH v5 03/22] sequencer: stop leaking buf Denton Liu
2020-04-07 14:27 ` [PATCH v5 04/22] sequencer: make file exists check more efficient Denton Liu
2020-04-08 0:10 ` Junio C Hamano
2020-04-07 14:27 ` [PATCH v5 05/22] sequencer: make read_oneliner() accept flags Denton Liu
2020-04-08 0:11 ` Junio C Hamano
2020-04-07 14:27 ` [PATCH v5 06/22] sequencer: configurably warn on non-existent files Denton Liu
2020-04-07 14:27 ` [PATCH v5 07/22] sequencer: make read_oneliner() extern Denton Liu
2020-04-07 14:27 ` [PATCH v5 08/22] rebase: use read_oneliner() Denton Liu
2020-04-08 0:26 ` Junio C Hamano
2020-04-07 14:27 ` [PATCH v5 09/22] sequencer: make apply_autostash() accept a path Denton Liu
2020-04-07 14:27 ` [PATCH v5 10/22] sequencer: rename stash_sha1 to stash_oid Denton Liu
2020-04-07 14:27 ` [PATCH v5 11/22] rebase: use apply_autostash() from sequencer.c Denton Liu
2020-04-07 14:27 ` [PATCH v5 12/22] rebase: generify reset_head() Denton Liu
2020-04-07 14:28 ` [PATCH v5 13/22] reset: extract reset_head() from rebase Denton Liu
2020-04-07 14:28 ` [PATCH v5 14/22] rebase: extract create_autostash() Denton Liu
2020-04-07 14:28 ` [PATCH v5 15/22] rebase: generify create_autostash() Denton Liu
2020-04-07 14:28 ` [PATCH v5 16/22] sequencer: extract perform_autostash() from rebase Denton Liu
2020-04-07 14:28 ` [PATCH v5 17/22] sequencer: unlink autostash in apply_autostash() Denton Liu
2020-04-07 14:28 ` [PATCH v5 18/22] sequencer: implement save_autostash() Denton Liu
2020-04-07 14:28 ` [PATCH v5 19/22] sequencer: implement apply_autostash_oid() Denton Liu
2020-04-07 14:28 ` [PATCH v5 20/22] merge: teach --autostash option Denton Liu
2020-04-07 14:28 ` [PATCH v5 21/22] t5520: make test_pull_autostash() accept expect_parent_num Denton Liu
2020-04-07 14:28 ` [PATCH v5 22/22] pull: pass --autostash to merge Denton Liu
2020-04-10 15:34 ` [PATCH v5 00/22] merge: learn --autostash Phillip Wood
2020-04-10 16:26 ` 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=194df5e2-4e5d-3150-d8a4-bb55bdef75be@gmail.com \
--to=phillip.wood123@gmail.com \
--cc=alban.gruin@gmail.com \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=johannes.schindelin@gmx.de \
--cc=liu.denton@gmail.com \
--cc=phillip.wood@dunelm.org.uk \
/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).