info-inventor
[Top] [All Lists]

Re: Problem with single-precision math functions and SbBasic.h on Linux

To: Alexandre Naaman <naaman@xxxxxxxxxxxxxxxxxxxx>
Subject: Re: Problem with single-precision math functions and SbBasic.h on Linux
From: Jean-Francois Panisset <panisset@xxxxxxxxxxxx>
Date: Sat, 19 Aug 2000 01:15:52 -0400
Cc: info-inventor@xxxxxxxxxxx, panisset@xxxxxxxxxxxx
In-reply-to: Your message of "Fri, 18 Aug 2000 21:57:54 PDT." <Pine.SGI.4.10.10008182152550.274413-100000@xxxxxxxxxxxxxxxxxxxx>
Sender: owner-inventor@xxxxxxxxxxx
 <Pine.SGI.4.10.10008182152550.274413-100000@xxxxxxxxxxxxxxxxxxxx>Alexandre Naa
man writes
>On Fri, 18 Aug 2000, Jean-Francois Panisset wrote:
>
>Salut Jean-Francois,
>
> Add -D_POSIX_SOURCE and -D_XOPEN_SOURCE to your compile line and it will
>compile just fine.

OK, that works too, although the rest of my stuff needs the single-precision
math functions, so I don't really want to compile with those options
(which end up getting rid of a lot of useful stuff in the includes).

>
>[clip]
>
>> /usr/include/bits/mathcalls.h:54: declaration of C function `float
>> acos(float)' conflicts with
>
>[clip]
>
>> 
>> In fact, Linux does have single-precision math functions, so this is
>> what is causing problems. In this case, the fix appears to be simple:
>> 
>> #if !defined(__sgi) && !defined(__linux)
>> 
>> seems to do the trick.
>
>Merci, I'll check it out this week-end. (note that some of these defines
>are found elsewhere in the .c++ files and should be dealt with too)
>
>A+,
>
>Alexandre.
>
>--
>Alexandre Naaman - naaman@xxxxxxx - La conformite est la mort de l'ame.
>

Thanks,
JF


Jean-Francois Panisset                                panisset@xxxxxxxxxxxx
Software Engineer
Discreet Logic 

<Prev in Thread] Current Thread [Next in Thread>