From: Phillip Wood <phillip.wood123@gmail.com>
Date: Fri, 09 Oct 2026 13:34:28 GMT
Subject: Re: [PATCH] branch: let --delete-merged find squash merged branches
Message-ID: <da5d6d28-b64b-4cfa-9998-93c69720b3c6@gmail.com>
In-Reply-To: <CALnO6CB2qPtv6Cr4LL1x=ZPmD_zLw1GzLUKgXE-NcTWObbiLkA@mail.gmail.com>

Hi Ben

On 08/10/2026 18:51, D. Ben Knoble wrote:
> On Sun, Oct 4, 2026 at 5:54 AM Phillip Wood <phillip.wood123@gmail.com> wrote:
>>
>> If we have
>>
>> (topic)  D - C - B - A
>>                         \
>>    (main)    M - Q - P - O -
>>               \          /
>>                 - - S - -
>>
>> where M is a squashed merge of topic I think we have
>>
>>       M^2^{tree} == topic^{tree}
>>       Merge-base(M^1, M^2) == Merge-base(topic, topic@{upstream})
>>       $(git rev-list --count --right-only M^1...M^2) == 1
> > Perhaps we are thinking of 2 different things? In practice when I see
> a merge created using GitHub's squash and merge option (and, I think,
> also when using `merge --squash`), there is no second parent. You
> could instead just get
> >      (main) S - Q - P - O
> > where S is A+B+C+D applied to Q (i.e., closer to a cherry-pick with
> --no-commit).

Oh, you're right, I'd misremembered how "merge --squash" worked.

> And we know that S is not necessarily tree-same to topic's D, so…?

I agree we need to use patch ids rather than trees in that case.

Thanks

Phillip