Bug#592420: osgIntrospection makes package 8x larger

2010-08-23 Thread Loic Dachary
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi, > Yes, disk is typically cheap these days; nevertheless, there is a > limit to how many packages can take up quite so much of it, and > library packages in particular should respect that limit. I share your concern regarding the package size. Alt

Bug#592420: osgIntrospection makes package 8x larger

2010-08-23 Thread Manuel A. Fernandez Montecelo
On Monday 23 August 2010 13:02:48 Cyril Brulebois wrote: > Manuel A. Fernandez Montecelo (10/08/2010): > > Splitting the package would be a reasonable thing to do from the > > point of view of the package size, but IMO it's not unreasonable for > > architectures where this is enabled to have packa

Bug#592420: osgIntrospection makes package 8x larger

2010-08-23 Thread Aaron M. Ucko
Alberto Luaces writes: > Cyril Brulebois writes: Thanks for re-including me in the discussion! > It was that light on amd64, but on the most popular platform, i386, it > has always been that large and nobody argued before. Check, for example, > lenny's version > > http://packages.debian.org/len

Bug#592420: osgIntrospection makes package 8x larger

2010-08-23 Thread Alberto Luaces
Hi, Cyril Brulebois writes: [...] >> In the meantime, I will be setting the priority to whishlist, if you >> don't have any objection, but it will remain open in the case that >> there's other people wanting to express their opinion. > > I'd call it a pretty serious bug for a library to eat tha

Bug#592420: osgIntrospection makes package 8x larger

2010-08-23 Thread Cyril Brulebois
Hi, I'm putting Aaron back in the recipients of the mail, the BTS still doesn't Cc submitters and people replying to bugs. Manuel A. Fernandez Montecelo (10/08/2010): > Splitting the package would be a reasonable thing to do from the > point of view of the package size, but IMO it's not unreason

Bug#592420: osgIntrospection makes package 8x larger

2010-08-09 Thread Manuel A. Fernandez Montecelo
Hi Aaron, I do think that you have a point, however as mentioned in #586256 there seems to be people in the upstream development mailing list suggesting that it should be enabled in general. We did so except for architectures where it would be unreasonable (e.g. armel, hppa, mips...), among ot