Debian Patches
Status for git-assembler/1.5-1
| Patch | Description | Author | Forwarded | Bugs | Origin | Last update |
|---|---|---|---|---|---|---|
| readme-rebase-onto-limitation.patch | doc: Put "rebase --onto" note in the rebase section. | Yuri D'Elia <wavexx@thregr.org> | not-needed | upstream, https://gitlab.com/wavexx/git-assembler/-/commit/a3044094ffa5c90bc8d3a473e2cd32f1112b3a77 | 2026-11-21 | |
| compact-relpath-valueerror.patch | fix: `compact_relpath`: fall back to `path` when `relpath` throws a `ValueError` | Erich Gubler <erichdongubler@gmail.com> | not-needed | upstream, https://gitlab.com/wavexx/git-assembler/-/commit/7e13f329283634378f5a9617637a10f49632ffdd | 2026-11-21 | |
| rebase-missing-fork-point.patch | fix: skip rebase when fork-point is missing but base is already an ancestor Git's get_fork_point() in commit.c walks every historical tip in the base reflog and returns nothing unless that set has exactly one merge-base with the topic, and that commit is itself one of those tips. See also git-merge-base(1), "Discussion on fork-point mode". A rewritten base reflog (rebase/reset leaving abandoned tips) trips this even when the current tip is already an ancestor. This can lead to an "assembly storm", where every time you assemble, the process starts from the top, and takes a while to get back to the point where there are actually conflicts to fix. Complement the fork-point check with an is_ancestor check, and mark the branch as up to date in that case, instead of marking the whole graph dirty. |
Etienne Samson <samson.etienne@gmail.com> | not-needed | upstream, https://gitlab.com/wavexx/git-assembler/-/commit/8183a1de3339186d3aeed33173311a6cefe4d3bb | 2026-11-21 |
All known versions for source package 'git-assembler'
- 1.5-1 (sid)
