configure: WARNING: sys/sysinfo.h: present but cannot be compiled

Torbjorn Granlund tg-please-reply-to-the-list-only at swox.com
Wed Oct 31 02:03:00 CET 2007


"John David Anglin" <dave at hiauly1.hia.nrc.ca> writes:

  I'm sorry that HP upset you but I have no connection with HP.
  
It is not so much about being upset, it is about having tons of code
stolen, and about not wanting to start over.

  I could provide alternative access to my personal HP hardware installed
  at NRC, HP-UX 11.11 and linux.  I'm also willing to test anything you come
  up with.  There are a number of linux machines online at HP outside the
  testdrive configuration that some open source developers have access to.
  Access to these machines could be arranged.
  
Thanks, but no thanks.  I am sure people will understand if my
entusiasm for working with the platform somewhat faltered.

  Since gmp is now required to build GCC, it is important that the
  parisc gmp port continue to work at a level sufficient to build GCC.
  Thus, I and others in the parisc community would appreciate your
  continued support.
  
If nobody else is willing to work on GMP/HP issues, I'd suggest that
the configuration used for HP should be "none", that should avoid most
trouble.  (I.e., use "configure --host=none".)

  This is a minor bug.  I was going to start looking into GCC PR 33584:
  
    GNU MP: Cannot reallocate memory (old_size=4 new_size=268435472)
    /xxx/gnu/gcc/gcc/gcc/testsuite/gfortran.dg/integer_exponentiation_4.f90:0:
    internal compiler error: Aborted
  
  At the moment, it's not clear where the problem lies.  However, if it
  is in GMP, it sounds like I shouldn't bother reporting it ;(
  
Please investigate it and send me a fix, if it'd turn up to be
GMP-related.  Probably the called asks for a too large number for the
VM available on the host system, which is of course not a GMP bug.)

-- 
Torbjörn


More information about the gmp-bugs mailing list