Re: [PATCH v3 6/7] rebase: write better reflog messages

From: Ramkumar Ramachandra <artagnon@gmail.com>
Date: Jun 24, 2013, 07:07 UTC
Message-ID: <CALkWK0kj3UOx8sq+h=L0giUC-vn+h3by9o_6YbjA8ArRXZfgZw@mail.gmail.com>
In-Reply-To: <7vfvw8dw9m.fsf@alter.siamese.dyndns.org>

Junio C Hamano wrote:

Show 10 quoted lines

>> @@ -59,6 +63,9 @@ else
>>               return $?
>>       fi
>>
>> +     # always reset GIT_REFLOG_ACTION before calling any external
>> +     # scripts; they have no idea about our base_reflog_action
>> +     GIT_REFLOG_ACTION="$base_reflog_action"
>>       git am $git_am_opt --rebasing --resolvemsg="$resolvemsg"
>
> Why does this reroll still use this base_reflog_action convention?

Because it's simple and it makes sense.

Show 5 quoted lines

> The original orig_reflog_action you borrowed this may have been an
> acceptable local solution inside
> git-rebase--interactive that does not call out to amyting, but
> the above comment a good demonstration that shows why this cannot be
> a good general solution that scales across scriptlets.

Nonsense. How do I set a custom reflog action when a certain flag is passed to my script (like git-rebase.sh:333) without *overriding* an existing GIT_REFLOG_ACTION? How do I construct cute start/pick/reword prefixes elegantly (like in git-rebase--interactive.sh:comment_for_reflog()) without *overriding* an existing GIT_REFLOG_ACTION? In both these examples, I'm setting a GIT_REFLOG_ACTION for the "rest of the code" to use. I don't care about the exact command sequence, but I know that they respect GIT_REFLOG_ACTION; so I'm setting one in advance.

When calling out to an external scriptlet, I want to define my own reflog message: when I call out to "am" from "rebase -i", it should write a "rebase -i: " message, and ignoring its own set_reflog_message(). _That_ can be done using the subshell thing you proposed. And I have absolutely no clue why

  (
    export GIT_REFLOG_ACTION
    git am
  )

is "more scalable" than

  GIT_REFLOG_ACTION="$base_reflog_action"
  git am

> And I already explained that to you at least twice.

You just gave set_reflog_action() and GIT_REFLOG_ACTION some sort of God status, and proposed to make the scripts more ugly and less extensible.

... and we're discussing absolutely trivial inconsequential rubbish once again. In any case, I've given up on arguing with you as it is clear that I can't possibly win. Do whatever you want.