I'm using gcc-config 1.3.14 with gcc-4.1.1 and wondered why gcjh doesn't function. Simple Solution: Add gcjh and gfortran, which also doesn't work to the list of binaries, patch follows
Created attachment 103725 [details, diff] Patch for gcc-config This fixes it for me (TM), haven't looked if there should be taken special care in the case of cross-compilation or so...
*** Bug 147087 has been marked as a duplicate of this bug. ***
Very good patch, but not enough... I was just switching from gcc-3.4.6 to 4.1.1 and started the emerge -e system/world process, but got a lot of fortran-related warnings / errors (libtool almost broke). But with your patch, plus the additional modifications I made to gcc-config, I was able to compile libtool without errors, and at this point emerge -e system is done - without errors - and emerge -e world is about 25% done - again without any serious, fortran-related errors. Thanks, without this lead I would probably still just sit and scratch my head... The updated patch follows.
Created attachment 103845 [details, diff] Updated gcc-config patch
Created attachment 103846 [details, diff] Updated gcc-config patch, v2 Oh, sorry, there was unnecessary cruft on the original patch. Here's a clean one.
(In reply to comment #5) > Created an attachment (id=103846) [edit] > Updated gcc-config patch, v2 > > Oh, sorry, there was unnecessary cruft on the original patch. Here's a clean > one. This patch may need to make it into v4.1.2 too. It fixed problems I had emerging various scientific packages like scipy with this gcc version.
this is basically waiting on the sci team to say "yeah this is OK"
Should be fine for gfortran ... can't speak for gcjh.
(In reply to comment #8) > Should be fine for gfortran ... can't speak for gcjh. > Works fine for gfortran on Linux 2.6.19-gentoo-r5 #7 SMP PREEMPT Sun Mar 11 22:50:09 GMT 2007 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 3800+ AuthenticAMD GNU/Linux Installing packages in R now works again, Thank you very much!
added to 1.3.15-r1 then, thanks for the input
(In reply to comment #10) > added to 1.3.15-r1 then, thanks for the input > The patch applied 1.3.15-r1 is a bit different from the previously submitted patch (which works for me). Unfortunately, for some reason 1.3.15-r1 does not work. Simple test: try "gfortran" with no argument It should return gfortran: no input files
fixed in cvs, it should be {,${CTARGET}-}{... not {,${CTARGET}}-{...