This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH 0/3] cond-optab merge
Paolo Bonzini wrote:
> Dave Korn wrote:
>> Paolo Bonzini wrote:
>>> Here is the final patch and results for cond-optab.
>> Hi Paolo, I haven't been following this in depth so apologies in advance if
>> this is just a misunderstanding; I would just like to ask for clarification:
>>
>>> test-cond* newlib compare bootstrap
>>> i386 almost no done
>>> The third column says whether newlib is supported on the target and can be
>>> built. The patch does not introduce new newlib build failures, and the ones
>>> in the baseline have been reported.
>> Newlib doesn't build on i386? Or is it just that it failed in your baseline
>> owing to some unrelated bug and so there's no change?
>
> Newlib builds on i386-linux, but it is not trivial to use it because
> sometimes files from the glibc installation get in the way, so I didn't
> test it. You have to build x86-linux apps with -nostdlib and manually
> specify newlib's crt0.o. Also, libgcov.a is built with glibc headers
> and it expects some magic symbols that aren't in newlib.
Right, thanks for the explanation. I'll run a testsuite on cygwin (which is
newlib-based) later this week after you've done the commit, which should
hopefully cover the newlib+i386 combination to at least some depth.
cheers,
DaveK