Re: daily pestering about s390 and sparc buildd failure
Thomas Bushnell BSG <tb@becket.net> writes:
> Ingo Juergensmann <ij@2004.bluespice.org> writes:
>
>> But - as you've figured out yourself - a buildd doesn't use pbuilder, but
>> sbuild. So, why haven't you tested with sbuild on your own?
>
> It's like you can't follow things for a moment. The bug would not and
> could not manifest itself on my arch.
>
>> How can you now that? A new uploads *always* triggers rebuilding on all
>> archs of your package.
>
> No, it doesn't. The new package hit the queues before the old one got
> built, and therefore the old one never got autobuilt (on the
> overloaded queues, that is).
But on the queues that just strugle to get along and have just build
or are still building your package.
Also note that the queue is not a FIFO. A new upload will cut in front
of other packages moving them back one position on archs where your
package has already been build. The further up in the queue sorting
your package is the more harmfull a new upload is (since more packages
below you are affected).
It is amazing how often Package builds get rejected or aborted due to
new uploads and for something like the mozilla clones with >50h build
time getting the build rejected realy sucks. Imagine someone uploading
mozilla again just to test if the s390 buildd has a bug.
MfG
Goswin
Reply to: