From: Yukihiro Matsumoto <matz@...>
Date: 2006-06-28T23:53:23+09:00
Subject: Re: 1.8.5p1 build failure on Solaris 10

Hi,

In message "Re: 1.8.5p1 build failure on Solaris 10"
    on Wed, 28 Jun 2006 23:37:00 +0900, ts <decoux@moulon.inra.fr> writes:

| Well I don't know if it's important but in configure.in there is
|
|solaris*2.10)  ac_cv_func_isinf=yes
|               LIBS="-lm $LIBS"
|              ;;

It is to fix [ruby-core:7138], that is

In message "[PATCH] solaris 10 isinf and ruby_setenv fixes"
    on Fri, 13 Jan 2006 21:56:10 +0900, <ville.mattila at stonesoft.com> writes:
>Hello,
>
> Here is a patch that fixes two issues for me.
> 1. The isinf is not regognized by autoconf library guesser, so it gets
>undefined and
>      the missing/isinf.c is used, which will get compilation error.
>      Solution is to force isinf to defined and -lm to be added to LIBS
>when solaris is configured.

If the patch from another Solaris user doesn't fix your problem,
something is different among Solaris2.10 platforms, and we have to
care about it.  Since I don't have a Solaris box, I am not sure
whether a fix should be pointing ininf() in -lm (or -lsunmath)
somehow, or modifying missing/isinf.c.

							matz.