On Wed, 2005-09-07 at 11:17 -0400, Daniel Veillard wrote:
On Wed, Sep 07, 2005 at 11:03:01AM -0400, Jeremy Katz wrote:
> On Wed, 2005-09-07 at 08:52 -0400, Daniel Veillard wrote:
> > In that case it's not a transcient failure, I think last time I tried to
> > use anaconda of rawhide back in July I had the exact same error.
>
> The error you're seeing is a "tree composed badly and broke".
That's
> unfortunately one that's dependent on the phase of the moon causing the
> build machines to be uncooperative. Hard to anticipate :)
So I was just unlucky on that too and the fact I get the same error
does not mean it was the state of the package in the meantime ? If yes
sorry about that I overgeneralized again.
Yeah, nothing to do with the package, just luck of the draw :)
Is detecting "tree composed badly and broke" possible ?
And avoid
pushing those ?
If we avoid pushing them, then no packages get pushed. Most of the
people following rawhide are just following the package updates, so that
seems like a less than ideal thing to do. And what if just one arch is
hosed (which is common). Ignoring the detection issue, which is hard at
best.
Jeremy