Re: [Monotone-devel] Re: revision.cc:166: invariant 'I(purported_m_child == m_child)' violated

[Nathaniel Smith]

On Sun, Jan 30, 2005 at 11:45:41AM +0100, Peter Simons wrote:
> Nathaniel Smith writes:
>  > Can you post --debug output?
> 
> http://peti.cryp.to/monobug-log

Okay, from looking at this, I think that it's just the bug in 0.16's
rebuild.  It created invalid root revisions.  Your database is shallow
enough that when 'commit' verifies history 3 commits back in time,
it's bumping into the root revision, and correctly pointing out that
it's invalid.

We'll be putting together some more systematic instructions on
rebuilding out of this particular issue when we release 0.17, but in
the mean time, using the latest head's 'db rebuild' _should_ fix
things.

>  > I strongly suspect "db rebuild" will fix things.
> 
> I tried that, but after the rebuild several of the branches
> behaves strangely (they were suddenly empty, for instance).

This worries me -- I don't know of any bugs that would cause that?
Let me know if it still happens...

-- Nathaniel

-- 
"But suppose I am not willing to claim that.  For in fact pianos
are heavy, and very few persons can carry a piano all by themselves."