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

Re: config.sub, upstream author responsibility?



>On POSIX, and especially Debian GNU/* you _shouldn't_ need to do
>architecture specific things. Sure, things like assembly optimization
>and tweaking compiler flags are nice, but they shouldn't be required
>for compilation. Compiling an unoptimized, untested version is better
>than the compilation failing just because there's a chance that it
>would fail if it tried.

Yeah.  Actually, this would work fine in most cases if config.sub just passed 
through a cpu name that it didn't recognise, rather than trying to act as 
arbiter of valid configurations.  (Virtually every program that cares about 
the cpu type will include its own check anyway.)  But instead config.sub 
currently errors out which means that the package has no target information at 
all to work with.

p.


Attachment: pgp7xoJvfI0d5.pgp
Description: PGP signature


Reply to: