Re: [PATCH v4 0/3] Use default values from settings instead of config

> If I comment yout your addition of GIT_TEST_COMMIT_GRAPH=0 in that file
> I see that we fail N number of tests, but all of them are actually
> fallout of just this test:
> 
>         git replace $head_parent $head && 
>         git replace -f $tree $blob 
> 
> I.e. we've created a replacement object replacing a tree with a blob, as
> part of tests I added to test how mktag handles those sorts of weird
> edge cases.
> 
> This then causes the graph verify code to throw a hissy fit with:
> 
>     root tree OID for commit 0ddfaf193ff13d6ab39b7cbd9eed645e3ee2f050 in
>     commit-graph is da5497437fd67ca928333aab79c4b4b55036ea66 !=
>     0fbca9850869684085d654f9e1380c9780802570
> 
> I.e. when we wrote the graph we somehow didn't notice that the root tree
> node we wrote is to an object that's not actually a tree? Isn't this a
> bug where some part of the commit graph writing should be doing its own
> extended OID lookup that's replacement-object aware, it didn't, and we
> wrote a corrupt graph as a result?
> 
> If there is a legitimate reason why we're not just hiding a bug we've
> turned up with these fixes let's disable that one test, not the entire
> test file.