Re: Is fetch.writeCommitGraph (and thus features.experimental) meant to work in the presence of shallow clones?

[Taylor Blau <me@ttaylorr.com>]

On Tue, Apr 14, 2020 at 05:50:57PM -0600, Taylor Blau wrote:
> On Tue, Apr 14, 2020 at 04:31:19PM -0400, Derrick Stolee wrote:
> > On 4/14/2020 4:22 PM, Elijah Newren wrote:
> > > Hi,
> > >
> > > I was building a version of git for internal use, and thought I'd try
> > > turning on features.experimental to get more testing of it.  The
> > > following test error in the testsuite scared me, though:
> > >
> > > t5537.9 (fetch --update-shallow):
> > >
> > > ...
> > > + git fetch --update-shallow ../shallow/.git refs/heads/*:refs/remotes/shallow/*
> > > remote: Enumerating objects: 18, done.
> > > remote: Counting objects: 100% (18/18), done.
> > > remote: Compressing objects: 100% (6/6), done.
> > > remote: Total 16 (delta 0), reused 6 (delta 0), pack-reused 0
> > > Unpacking objects: 100% (16/16), 1.16 KiB | 1.17 MiB/s, done.
> > > From ../shallow/
> > >  * [new branch]      master     -> shallow/master
> > >  * [new tag]         heavy-tag  -> heavy-tag
> > >  * [new tag]         light-tag  -> light-tag
> > > error: Could not read ac67d3021b4319951fb176469d7732e6914530c5
> > > error: Could not read ac67d3021b4319951fb176469d7732e6914530c5
> > > error: Could not read ac67d3021b4319951fb176469d7732e6914530c5
> > > fatal: unable to parse commit ac67d3021b4319951fb176469d7732e6914530c5
> > >
> > > Passing -c fetch.writeCommitGraph=false to the fetch command in that
> > > test makes it pass.
> > >
> > > There were also a couple other tests that failed with
> > > features.experimental=true (in t5500), but those weren't scary -- they
> > > were just checking exact want/have lines and features.experimental is
> > > intended to change those.  This test from t5537 was the only one that
> > > showed some unexpected fatal error.
> >
> > Well, commit-graphs are not supposed to do anything if we have
> > shallow clones. We definitely don't load a commit-graph in that
> > case. Seems like we need an extra check in write_commit_graph()
> > to stop writing in the presence of shallow commits.
>
> This rang a bell to me, too. There's a bug, but it's due to the mutative
> side-effects of 'is_repository_shallow' along with '--update-shallow' (a
> normal 'git fetch' works fine here, with or without
> fetch.writeCommitGraph).
>
> Here's a patch that I didn't sign-off on that fixes the problem for me.

Oh, apparently I did sign this one off ;). I'll blame that 'git commit
--amend -vs' is muscle memory for me. In either case, this probably
needs some work (if it's even the right approach) before queueing it.

> --- >8 ---
>
> ...

Thanks,
Taylor