On 9 June 2025 3:13:01 am IST, "brian m. carlson" <sandals@xxxxxxxxxxxxxxxxxxxx> wrote: >On 2025-06-08 at 19:35:24, Aditya Garg wrote: >> >> >> On 9 June 2025 1:01:19 am IST, Phillip Wood <phillip.wood123@xxxxxxxxx> wrote: >> >On 08/06/2025 19:57, brian m. carlson wrote: >> >> >> >> If your goal is to just do the fixups and squash and not anything else >> >> interactive, then you can do this: >> >> >> >> GIT_SEQUENCE_EDITOR=true git rebase -i --autosquash >> > >> >Just a quick note to say that since git v2.44.0 you can just run >> > >> > git rebase --autosquash >> > >> >> Silly question but how does it get to know what is the fixup commit? > >That's actually a great question. When you do `git commit --squash >REVISION`, it takes the summary of the commit specified by `REVISION` >and precedes it with `squash! `, and for `--fixup`, it does `fixup! `. > >Then, it's just a matter of re-ordering the squash or fixup commits in >order after the commit with the corresponding summary. > >So with this shell script: > >---- >#!/bin/sh > >git init --object-format=sha256 test-repo >cd test-repo > >echo abc >file.txt >git add file.txt >git commit -m 'Initial commit' > >echo def >file.txt >git add file.txt >git commit --fixup HEAD >git show >---- > >You get something like this: > >---- >Initialized empty Git repository in /tmp/user/1000/test-repo/.git/ >[dev (root-commit) 7327102] Initial commit > 1 file changed, 1 insertion(+) > create mode 100644 file.txt >[dev 8bdd271] fixup! Initial commit > 1 file changed, 1 insertion(+), 1 deletion(-) >commit 8bdd271b6d4e22b7ca697c2d4499fd3e0825977d7d2c917b92e1f1f12383f52c >Author: brian m. carlson <sandals@xxxxxxxxxxxxxxxxxxxx> >Date: Sun Jun 8 21:41:43 2025 +0000 > > fixup! Initial commit > >diff --git a/file.txt b/file.txt >index e0ef420..559afde 100644 >--- a/file.txt >+++ b/file.txt >@@ -1 +1 @@ >-abc >+def >---- > >Of course, `--autosquash` does require using `--fixup` and `--squash` or >otherwise naming the commits that way. Now it's clear. Thanks a lot brian and Junio!
