Re: rebase -i reword runs pre-commit hook with curious results

From: Neal Kreitzinger <nkreitzinger@gmail.com>
Date: Feb 2, 2012, 16:39 UTC
Message-ID: <4F2ABC5B.2030608@gmail.com>
In-Reply-To: <4F2A2286.3090808@sohovfx.com>

On 2/1/2012 11:43 PM, Andrew Wong wrote:

Show 13 quoted lines

> On 12-02-01 4:50 PM, Neal Kreitzinger wrote:
>> I'm confused on why and/or how interactive rebase runs the pre-commit
>> hook
>> when doing the reword command for commit (a).
> When you do a "reword" in "rebase -i", it basically does a "cherry-pick"
> of that commit first, then it does a "commit --amend". And your
> pre-commit hook should've been run during the amend.
>> IOW, the pre-commit hook does not get the same results as if I were
>> doing a
>> commandline git-commit of a modified index.
> Does your pre-commit hook work when doing a "commit --amend"? I'm not
> sure if you can actually modify the author (or committer) date from
> inside a pre-commit hook.

(We have a comment on "line 1" in our source with $User:$ $Date:$ keywords that the pre-commit hooks expands to insert "whoami" and "date" values to effect a user-datestamp at commit time. We do this to enforce conflicts on same-file edits.) Now that I understand that the cherry-pick takes place first to effect the transfer of the tree content and then a subsequent git-commit --amend of "no changes" takes place to effect the reword opportunity, the behavior makes sense now. (We use git-commit --amend to reword commit messages also.) The pre-commit hook runs prior to commit message editor just like commandline git-commit --amend (and plain git-commit).

thanks!

v/r, neal