On 2024-10-16 10:57, Patrick Steinhardt wrote: > Given that we do set `log.abbrev` I think we should be hitting code > paths in git-shortlog(1) that use it. `git shortlog --format=%h` for > example would use `log.abbrev`, wouldn't it? It would be nice to figure > out whether this can be made to misbehave based on which object hash we > have in the input. I did try this, but like I said, --format seems not supported when git-shortlog(1) is reading from stdin. It always outputs the same summary format, either grouped on authors or committers. This is not explicitly documented anywhere - the manual only says that "git shortlog will output a summary of the log read from standard input" and then goes on to document all the options with no mention of a difference in behaviour when reading from stdin. So I'm still not entirely convinced that this is impossible to trigger (also given the complexity of the argument parsing machinery), but I have not been able to find a way. > I think it doesn't hurt to have an explicit test for this scenario, even > if it just demonstrates that things don't crash or behave in weird ways. Sure, that makes sense. > You can take e.g. b2dbf97f47 (builtin/index-pack: fix segfaults when > running outside of a repo, 2024-09-04) as inspiration for how to achieve > this. Thanks much for the pointer, I'll have a look! -- Wolf
