[v3] Handle different versions of Solaris 8 <iso/math_iso.h>, <iso/stdlib_iso.h>

[Paolo Carlini]

Hi,
> Hi Paolo,
>
>>> I guess the patch is ok now?
>> Yes. Nice that Jon replied in the meanwhile and clarified the undefined behavior issue: for the time being I think we can keep the __cplusplus checks, should also help during this testing period. We can certainly clean up the thing later. Maybe you could add a comment somewhere summarizing what Jon wrote.
> for the moment, I've installed the patch as is.  It's probably better to
> add some statement along those lines to user docs (and mention the
> change in gcc-4.7/changes.html :-)
Ok.
> Btw., I just noticed the following in Oracle Studio 12.3 Beta CC(1):
>
>       -features=a
>            Enables/disables various C++ language features.
>
>            The following table lists the -features suboption key-
>            words and their meanings.  The prefix no% applied to a
>            suboption disables that suboption.
>
>            Value          Meaning
> [...]
>            cplusplus_redef
>                 Allow the normally pre-defined macro __cplusplus
>                 to be redefined by a -D option on the command
>                 line. Attempting to redefine __cplusplus with a
>                 #define directive in source code is not allowed.
>
>                 Example:
>                  CC -features=cplusplus_redef -D__cplusplus=1 ...
>
>                 The g++ compiler typically predefines the
>                 __cplusplus macro to 1, and source code might
>                 depend on this non-standard value. (The standard
>                 value is 199711L for compilers implementing the
>                 1998 C++ standard or the 2003 update. Future stan-
>                 dards will require a larger value for the macro.)
>
>                 Do not use this option unless you need to redefine
>                 __cplusplus to 1 in order to compile code intended
>                 for g++.

Interesting. I suppose you are going to inform the SunStudio engineer 
that some of the above has to be adjusted for gcc 4.7+ (otherwise, 
please let me know, I can probably do it via Stephen Clamage)

Thanks,
Paolo.