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

Re: Canonical and Debian

On Tue, Jun 07, 2005 at 10:31:24PM +0200, Tollef Fog Heen wrote:
> * Julien BLACHE 
> | A bug is a bug, whether it triggers or not.
> It's not RC and therefore not a priority if it has no effect.

The primary question, I think, is whether one can be 100% sure whether a
bug that results in an FTBFS on only one out of eleven platforms will
not have any effect whatsoever on another platform.

Usually, the answer to that one is "no, you can't be sure". FTBFS bugs
that occur on only one platform are rare, very rare; most build failures
are mistakes in packaging (which usually have effect on all
architectures, rather than just one) or things such as incorrect
assumptions regarding char signedness or word length, that have effect
on all big endian or 64-bit platforms. Of course, these usually result
in runtime errors rather than compile time ones.

The one rare problem which /does/ affect only one platform is either a
toolchain bug (which is not a bug in the package failing to build,
although workarounds would need to be applied to that package) or a
complete and severe fuckup of the type that only exist because of
braindamaged or malicious upstream authors. I don't think I want that
kind of code on my system, even if that one bug does not affect my
architecture. As said, though, these are rare, very rare.

Of course I'm not talking about i386-pc-gnu and i386-pc-kfreebsd-gnu
here :-)

> Similarly, I'd expect Debian maintainers to care less about bugs which
> only affects OpenBSD or Windows than those which affect Linux on i386.

Oh, yes, but then there's a difference; Windows isn't POSIX compliant
(so you could need many #ifdefs), and OpenBSD is deliberately different
in some areas (such as the missing support for AF_INET/AF_INET6 hybrid

The amount of time between slipping on the peel and landing on the
pavement is precisely one bananosecond

Reply to: