Re: [oe] [RFC] Layers, PRINC and bbappends

[Richard Purdie <richard.purdie@linuxfoundation.org>]

On Thu, 2013-05-16 at 11:29 +0200, Martin Jansa wrote:
> On Thu, May 16, 2013 at 10:10:45AM +0100, Paul Eggleton wrote:
> > On Thursday 16 May 2013 10:59:54 Martin Jansa wrote:
> > > On Thu, May 16, 2013 at 09:29:48AM +0100, Paul Eggleton wrote:
> > > > On Thursday 16 May 2013 08:31:33 Koen Kooi wrote:
> > > > > Removing layers is pretty much impossible, though.
> > > > 
> > > > With PRINC gone would there be any other issue that would make removing
> > > > layers a problem?
> > > 
> > > Does PR service really fix removing layers? I don't think so, if you add
> > > layer and checksum of foo is changed AUTOPR is incremented, but then you
> > > remove the layer and the checksum is unfortunately the same as what you
> > > had before AUTOPR will go backwards.
> > 
> > Then I would have to say if that's not the behaviour you want we should change 
> > it or at least make it optional to force it to always increment even if the 
> > checksum is the same as a previous value.
> 
> That won't work with multiple builders using the same PR server (I don't
> wont other builders to always increment PRAUTOs used in builds which
> populate official feed).

I just went and looked at the code and as far as I can tell PRServ
defaults to always incrementing upon a change. To get the other
behaviour you'd need nohist=False to PRData() in lib/prserv/db.py and I
can't see where that is set other than to the default :/. I'm tired and
perhaps missing something though.

Cheers,

Richard