Re: Autobuilding with cross-compiler to solve m68k woes?
Hi Sam,
> Steve> What if we had the means of reliably determining which
> Steve> packages could be cross-compiled, and which needed to be
> Steve> compiled on a native system? Would that make
> Steve> cross-compiling a viable option?
> Please go successfully cross compile multiple Debian packages you do
> not maintain. You probably don't want to upload the results but if
> you successfully generate debs from a Debian source package by cross
> compiling, it will be an educational experience.
> Those of us who have played with cross compilers tend to believe that
> the engineering problems will dominate and that it is not a viable
> solution.
> If you do have experience making cross compilation work for large
> systems, then please start out with a more fleshed out proposal that
> demonstrates you have thought about the hard problems.
'We tend to believe that the engineering problems will dominate and that it is
not a viable solution' sounds a lot different to me than 'cross-compiling is
evil'. :) I recognize that there are many technical issues here that would need
to be resolved before cross-compiling would be a large-scale option, not least
of which to my eye is the need for a complete, prepackaged cross-compiling
toolchain that can be rebuilt from source with a minimum of pain. I'm trying to
look long-term here; and maybe it's naivété that makes me say the technical
issues are not insurmountable, so I would like to know what the technical issues
are that have convinced others that it will never be workable.
No, I don't have a fleshed-out proposal yet; I still need to go through a few
more 'educational experiences' first, as you say. But I'm optimistic that down
the line, cross-compilers might be easy enough to set up that they'll be useful
to Debian as a whole.
Steve Langasek
postmodern programmer
Reply to: