Re: The 98% and N<=2 criteria
Henning Makholm wrote:
> Scripsit David Weinehall <email@example.com>
> > That said, I'm a firm believer of the suggestion posed by Jesus
> > Climent, that we should have base set of software (where base is
> > probably a bit bigger than our current base) released for all
> > architectures that have a working installer, and then only have full
> > official releases for a limited set of architectures.
> Such a base set of software would surely include a compiler toolchain,
> wouldn't it? If sounds plausible that the toolchain is the collection
> of software where architecture-specific bugs are _most_ likely to turn
> up, so would we actually have gained anything then?
Toolchain bugs which affect the toolchain itself are usually quickly
resolved, once a port is somewhat stable. If this isn't the case, then
the port can't even stay in unstable for long.