Re: [PATCH 2/3] rebase: help users when dying with `preserve-merges`

From: Junio C Hamano <gitster@pobox.com>
To: Philip Oakley <philipoakley@iee.email>
Cc: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>,
	"Philip Oakley via GitGitGadget" <gitgitgadget@gmail.com>,
	git@vger.kernel.org
Subject: Re: [PATCH 2/3] rebase: help users when dying with `preserve-merges`
Date: Thu, 26 May 2022 13:42:19 -0700	[thread overview]
Message-ID: <xmqq1qwgxbys.fsf@gitster.g> (raw)
In-Reply-To: <00229772-f075-0b0c-7810-7debf6b971bc@iee.email> (Philip Oakley's message of "Thu, 26 May 2022 12:44:49 +0100")

Philip Oakley <philipoakley@iee.email> writes:

>>> Make the `rebase --abort` option available to allow users to remove
>>> traces of any preserve-merges rebase, even if they had upgraded
>>> during a rebase.

This patch does not make it "available", though.  

	Suggest using `--abort` to get out of the situation after a
	failed preserve-rebase and remove traces of ...

perhaps?

I do think the suggestion is worth doing if a user ever gets into
the situation, but how likely does it happen?  A user has to start
"rebase -p" with older Git, wait until Git gets updated to a future
version of Git that includes this change, and then say "rebase -p
--continue"?

>>>  	} else if (is_directory(merge_dir())) {
>>>  		strbuf_reset(&buf);
>>>  		strbuf_addf(&buf, "%s/rewritten", merge_dir());
>>> -		if (is_directory(buf.buf)) {
>>> -			die("`rebase -p` is no longer supported");
>>> +		if (is_directory(buf.buf) && !(action == ACTION_ABORT)) {
>>> +			die("`rebase --preserve-merges` (-p) is no longer supported.\n"
>>> +			"Use `git rebase --abort` to terminate current rebase.\n"
>>> +			"Or downgrade to v2.33, or earlier, to complete the rebase.\n");
>>>  		} else {
>>>  			strbuf_reset(&buf);
>>>  			strbuf_addf(&buf, "%s/interactive", merge_dir());
>> Existing issue: No _(), shouldn't we add it?
> This `strbuf_addf` is forming a path for internal use. It just happens
> to look like legible English ;-)

I do not think Ævar meant "%s/interactive"; the enhanced message
above that you inherited from the original "no longer supported"
that was not marked for translation.

>> I wonder if we should use die_message() + advise() in these cases,
>> i.e. stick to why we died in die_message() and have the advise() make
>> suggestions, as e4921d877ab (tracking branches: add advice to ambiguous
>> refspec error, 2022-04-01) does.
>
> Ah, maybe it's my message.. that needs translating.

Yup.

This whole '-p' business will go away in a few releases down, so a
longer message give to the existing die() should be sufficient and
there is no need for the choice between "yes, I am still weaning
myself off of rebase -p and want to keep seeing the advice" and
"thanks, I saw the message often enough, you no longer need to tell
me how to get out", I would think.

Thanks.