[Gnu-arch-users] Re: patch: automatic remove the corrupt revlib revision

[Stefan Monnier]

> That's a big assumption, isn't it?  Specifically, is "a strong hash of
> the whole revision" significantly easier to compute reliably than a
> Git tree object?

What's so hard about computing a hash of a whole revision? especially
knowing this is a revision in a revlib, i.e. there should be no extra
spurious file, except for the ones tla puts there.


        Stefan