Re: ARMv4-support in armel/squeeze?

[Paul Brook <paul@codesourcery.com>]

>  you have to bear in mind that "future porting" *used* to be very
> uncommon an occurrence (as if doing it 14 times is "uncommon")
> 
>  but with the massive explosion in compiler options for ARM processors
> alone, the process of "porting" now becomes a massive headache.

Really? AFAIK the only really interesting option (soft vs. hard float ABI) has 
already been done.  I guess we might want a big-endian port eventually, but 
all the other options are just optimizations and fully binary compatible with 
the current ports. i.e. you can start from a full armel/armhf system and 
incrementally rebuild optimized packages from there. No reason to bootstrap 
from scratch.  The real problem is that dpkg can't handle multiple variants of 
the same package for the same architecture.  Worst case you invent a "new" 
port and --force-architecture install the armel packages.

Cross-compiling a base system from scratch may still be an interesting 
facility. However if you're doing that then IMO it's also reasonable to want 
to cross compile everything.  I don't believe bootstrapping of new ports is 
sufficient justification for this.  I'd expect that any vaguely automated 
system for such builds would bitrot rapidly, and require significant 
maintenance.

Paul