[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RFS: frown

Op za, 16-09-2006 te 18:20 +0100, schreef James Westby:
> > > I have a couple of questions if you could answer them for me, (purely
> > > out of interest 
> > > You have gone with cdbs handling debian/control. This is unpopular, can
> > > you tell me why you chose it?
> > Well it is very bad manners to do this on build time and the ftp-masters
> > will reject the package if it does this. 
> Yes I realise that. Though I don't think a REJECTED mail is strong
> enough for this!
Well the reason for that are quite clear in the REJECT FAQ, which boils
down to that it is against policy and that is should be clear which
binary packages a package produces by looking at the debian/control

Furthermore updating the control file during build can cause changes in
the source package between builds which is a side-effect that should be
> > > The update-haskell-control stuff seems unecessary to me (I realise it is
> > > not your choice.) Can you tell me why it is done like this?
> > Well it substitutes some Haskell related variables (right build
> > dependencies on ghc6, list of ghc6 arches etc) which can be especially
> > usefull if the package contains a Haskell library. But right now I only
> > use it to keep track of the list of Architectures frown can be build
> > on. 
> I was more asking why it added a list of architectures. Almost every
> other package gets by with arch: any. It is also recommended to only
> restrict the architectures for very specific reasons, does that apply to
> Haskell?
You could make a case for making it Arch: any, but I decided to keep
using the list provided by update-haskell-control to keep all my Haskell
packages consistent. The list of architectures that support ghc6 does
not change that quickly so it is no problem in this case.

I guess it is a matter of taste here and there is not big technical
reason why I choose the specific architecture list instead of the usual
Arch: any.

Greetings Arjan 

Attachment: signature.asc
Description: Dit berichtdeel is digitaal ondertekend

Reply to: