Skip to contentSkip to conversation

Bug report: git remote prune does not take negative refspecs into account

DISCUSSION Christian Wolf 1 participant 1 message Feb 14, 2021

lore
Share & tools
lore
#0 lore raw

Hello,

I found an issue with git and wanted to report this. Here is the filled result from git bugreport (sorry for the intermixed German, but I think the English statements should suffice to understand the issue).

Vielen Dank für das Ausfüllen eines Git-Fehlerberichts!
Bitte antworten Sie auf die folgenden Fragen, um uns dabei zu helfen, Ihr Problem zu verstehen.

Was haben Sie gemacht, bevor der Fehler auftrat? (Schritte, um Ihr Fehler zu reproduzieren)

1. I have configured a remote upstream with a refspec like +refs/heads/ *:refs/upstream/heads/*.
2. I have fetched remote upstream fetching two branches foo and bar.
3. I add another line to .git/config with a negative refspec

```
fetch = ^refs/heads/bar
```

4. I called git fetch upstream and git remote prune upstream.
5. Remove the refspec from step 2 in .git/config with some arbitrary names (e.g. +refs/heads/baz:refs/upstream/heads/baz), keep the negative refspec as it is
6. Fetch and prune (see step 4)
7. Undo step 5 and restore the original state as after step 3
8. Fetch and prune (see step 4)

Was haben Sie erwartet, was passieren soll? (Erwartetes Verhalten) Expected behavior:

After the fetch in step 2 both branches are mirrored locally.

After the prune in line 4 the branch bar is dropped from the remotely fetched branches.

After step 6 both both remote branches foo and bar are no longer visible. A possible error about unmatching branch baz.

After step 8 again a remote branch foo is existing.

Was ist stattdessen passiert? (Wirkliches Verhalten) Real behavior:

After step 2: As intended

After step 4: Both remote branches are still visible. Nothing is pruned.

After step 6: Both remote branches are gone.

After step 8: Only foo is existing/fetched as remote branch.

Was ist der Unterschied zwischen dem, was Sie erwartet haben und was wirklich passiert ist?
The difference of real and ideal behavior:

The difference is that after step 4 the remote branch bar is not pruned.

Sonstige Anmerkungen, die Sie hinzufügen möchten:

The steps 5~8 are there to verify the negative pathspec is working in general. Fetching freshly takes the negativbe pathspec into account.

So the problem is within the pruning code.

[System Info]
git Version:
git version 2.30.1
cpu: x86_64
no commit associated with this build
sizeof-long: 8
sizeof-size_t: 8
shell-path: /bin/sh
uname: Linux 5.10.10-arch1-1 #1 SMP PREEMPT Sat, 23 Jan 2021 23:59:48 +0000 x86_64
Compiler Info: gnuc: 10.2
libc Info: glibc: 2.33
$SHELL (typically, interactive shell): /bin/bash

[Aktivierte Hooks]

Thank you for looking into this.
Christian