[Piuparts-devel] RFC: preview/shuffle-slave-scripts

[Holger Levsen]

Hi,

On Sonntag, 20. Januar 2013, Andreas Beckmann wrote:
> > one question: why?
> because the master has nothing to do with the slave, for (a) remote
> slave(s) it does not even know where this/they reside(s) ...

right, haven't though much about remote slaves.
 
> > And I also like the multiple slave goal for 0.50! piatti will move to a
> currently working on it ...

yay!
 
> > virtual machine in future and then I would like to start running an i386
> > slave as well, to test i386 only packages.
> So one master and two architectures, somehow "combined"? That will bring
> a few interesting challenges ...

yes :)

> (trivial would be [sid-amd64] and [sid-i386] as independent full distro
> tests)

yup. but rather useless too. though maybe it would help detect failures due to 
locally built (in a dirty environment) uploads. _maybe_.. :)

> BTW, there are some arch:all packages with dependency-does-not-exist
> there the dependecy is arch:i386 only

eek. but we dont need piuparts to find these problems.


cheers,
	Holger