Re: [RFC PATCH] builtin/shortlog: explicitly set hash algo when there is no repo

[Taylor Blau <me@ttaylorr.com>]

On Tue, Oct 15, 2024 at 03:54:51PM -0400, Taylor Blau wrote:
> On Tue, Oct 15, 2024 at 11:33:35AM +0200, Wolfgang Müller wrote:
> > On 2024-10-11 20:34, Wolfgang Müller wrote:
> > > Whilst git-shortlog(1) does not explicitly need any repository
> > > information when run without reference to one, it still parses some of
> > > its arguments with parse_revision_opt() which assumes that the hash
> > > algorithm is set. However, in c8aed5e8da (repository: stop setting SHA1
> > > as the default object hash, 2024-05-07) we stopped setting up a default
> > > hash algorithm and instead require commands to set it up explicitly.
> > >
> > > This was done for most other commands like in ab274909d4 (builtin/diff:
> > > explicitly set hash algo when there is no repo, 2024-05-07) but was
> > > missed for builtin/shortlog, making git-shortlog(1) segfault outside of
> > > a repository when given arguments like --author that trigger a call to
> > > parse_revision_opt().
> > >
> > > Fix this for now by explicitly setting the hash algorithm to SHA1.
> > >
> > > Signed-off-by: Wolfgang Müller <wolf@oriole.systems>
> > > ---
> > > I think there is a much cleaner patch here if one would look at
> > > disentangling the parsing machinery split between cmd_shortlog() and
> > > parse_revision_opt() such that the latter is only called if there is an
> > > actual repository, but I'm not at all familiar enough with the codebase
> > > to do that. Like the commit says some other commands were fixed like
> > > this as well, so I thought to send this patch your way.
> >
> > Was this topic maybe missed? It's not a critical fix but would solve a
> > segfault that is pretty easy for a user to trigger.
>
> Yeah. It looks like this was sent out on Friday, which was Junio's last
> working day before vacation. I think we probably both assumed the other
> picked it up ;-).

I picked this up today and applied it to 'seen', but ejected it before
pushing the result out, as this appears to break CI, presumably due to
the mixed declaration and code in the patch.

Thanks,
Taylor