Re: [PATCH] gnu: cross-gcc-arguments: Disable libitm, libvtv and, libsanitizer.

[ludo@gnu.org (Ludovic Courtès)]

Manolis Ragkousis <manolis837@gmail.com> skribis:

> This is a follow up to better explain what the problem is with the
> sans-libc cross gcc in the (gnu packages cross-base) module.
>
> In the latest master if you try to build gcc-cross-sans-libc with:
>
> ./pre-inst-env guix build -e '((@@(gnu packages cross-base) cross-gcc)
> "i686-linux-gnu")' -c8
>
> gcc modules libitm, libvtv and libsanitizer fail to build with
> "configure: error: C compiler cannot create executables".

What I don’t get is that I have:

--8<---------------cut here---------------start------------->8---
$ git describe 
v0.9.0-2029-g863c4ee
$ ./pre-inst-env guix build -e '((@@ (gnu packages cross-base) cross-gcc) "mips64el-linux-gnu")' -n
The following files would be downloaded:
   /gnu/store/13nnp704cvxh05gxkbpqvjvhvqwf6b24-gcc-cross-sans-libc-mips64el-linux-gnu-4.9.3
   /gnu/store/bgm8y2j42dmviyc51kk2wcgjhvfpmc08-binutils-cross-mips64el-linux-gnu-2.25.1
   /gnu/store/d9rk59ff8zc42726fr2zdr572kb152i6-isl-0.11.1
   /gnu/store/hm1mw6lxdiaxq0370rigbwspj31nn1a8-mpc-1.0.3
   /gnu/store/sbsrma5yd943smdvcr7wnd2x5hd8jm7c-cloog-0.18.0
   /gnu/store/vachrk9j3f8yrh52v1yiwjv40dhaahnh-ld-wrapper-mips64el-linux-gnu-0
$ ./pre-inst-env guix build -e '((@@ (gnu packages cross-base) cross-gcc) "mips64el-linux-gnuabi64")' -n --no-grafts
substitute: updating list of substitutes from 'http://hydra.gnunet.org'... 100.0%
The following files would be downloaded:
   /gnu/store/w97wggnw4vqzknlnvhplfm722spsh66r-gcc-cross-sans-libc-mips64el-linux-gnuabi64-4.9.3
   /gnu/store/54cam4jxh59kv32xawq531h2hz397bia-binutils-cross-mips64el-linux-gnuabi64-2.25.1
   /gnu/store/5b2k4mmfyp3vimjqiq8msk2dsyaj2yg4-ld-wrapper-mips64el-linux-gnuabi64-0
   /gnu/store/d9rk59ff8zc42726fr2zdr572kb152i6-isl-0.11.1
   /gnu/store/hm1mw6lxdiaxq0370rigbwspj31nn1a8-mpc-1.0.3
   /gnu/store/sbsrma5yd943smdvcr7wnd2x5hd8jm7c-cloog-0.18.0
--8<---------------cut here---------------end--------------->8---

So maybe the failure is specific to i686 or something?

(In which case it’s of course still worth fixing.  Just trying to
understand.)

Thanks for the analysis, and apologies for the loooong delay!

Ludo’.