I'm compiling GDC on my FreeBSD machine now, and it seems to be working OK.  One caveat:  FreeBSD (and I suspect all BSDs) need to be detected the same as Mac OS X in terms of nan.h et al.  I'm pretty sure that that arrangement of headers as well as the lack of an fmodl is common to all systems with the BSD LIBC.

Owen

Well, the compiler seems OK, but compiling Phobos fails:

gdc -o gcc/config.o -g -O2 -I ../../gcc-3.3.3/gcc/d/phobos -I
./../gcc-3.3.3/gcc/d/phobos/internal/gc -c gcc/config.d
./../gcc-3.3.3/gcc/d/phobos/std/c/stdio.d(239): identifier 'Csnprintf' of
'gcc.config.Csnprintf' is not defined
gmake: *** [gcc/config.o] Error 1


In article <c3odae$2dsu$1@digitaldaemon.com>, resistor@mac.com says...
>
>I'm compiling GDC on my FreeBSD machine now, and it seems to be working OK.  One caveat:  FreeBSD (and I suspect all BSDs) need to be detected the same as Mac OS X in terms of nan.h et al.  I'm pretty sure that that arrangement of headers as well as the lack of an fmodl is common to all systems with the BSD LIBC.
>
>Owen
>
>

The error is due to a lack for FreeBSD-specific code. See my message for "release 1c".  For now, you could try copying the darwin-specific portions.

Does FreeBSD not have any long double math functions or is just fmodl missing?

In article <c3odnu$2en2$1@digitaldaemon.com>, "resistor" <resistor@mac.com> wrote:

> Well, the compiler seems OK, but compiling Phobos fails:
> 
> gdc -o gcc/config.o -g -O2 -I ../../gcc-3.3.3/gcc/d/phobos -I ./../gcc-3.3.3/gcc/d/phobos/internal/gc -c gcc/config.d ./../gcc-3.3.3/gcc/d/phobos/std/c/stdio.d(239): identifier 'Csnprintf' of 'gcc.config.Csnprintf' is not defined gmake: *** [gcc/config.o] Error 1
> 
> 
> In article <c3odae$2dsu$1@digitaldaemon.com>, resistor@mac.com says...
>>
>>I'm compiling GDC on my FreeBSD machine now, and it seems to be working OK.  One caveat:  FreeBSD (and I suspect all BSDs) need to be detected the same as Mac OS X in terms of nan.h et al.  I'm pretty sure that that arrangement of headers as well as the lack of an fmodl is common to all systems with the BSD LIBC.
>>
>>Owen
>>
>>
>>
>

http://www.dinkumware.com/conform_c.html#freebsd_results

There's a large number of the optional C95 math ops that are not implemented in the FreeBSD libc.  The above link has conformance test information about it. Note that OS X uses the FreeBSD libc, so any problems with it will affect OS X (and I suspect all the other BSDs as well).

Owen

In article <20040323.173001.496987743.15787@earthlink.net>, David Friedman says...
>
>The error is due to a lack for FreeBSD-specific code. See my message for "release 1c".  For now, you could try copying the darwin-specific portions.
>
>Does FreeBSD not have any long double math functions or is just fmodl missing?
>
>In article <c3odnu$2en2$1@digitaldaemon.com>, "resistor" <resistor@mac.com> wrote:
>
>> Well, the compiler seems OK, but compiling Phobos fails:
>> 
>> gdc -o gcc/config.o -g -O2 -I ../../gcc-3.3.3/gcc/d/phobos -I ./../gcc-3.3.3/gcc/d/phobos/internal/gc -c gcc/config.d ./../gcc-3.3.3/gcc/d/phobos/std/c/stdio.d(239): identifier 'Csnprintf' of 'gcc.config.Csnprintf' is not defined gmake: *** [gcc/config.o] Error 1
>> 
>> 
>> In article <c3odae$2dsu$1@digitaldaemon.com>, resistor@mac.com says...
>>>
>>>I'm compiling GDC on my FreeBSD machine now, and it seems to be working OK.  One caveat:  FreeBSD (and I suspect all BSDs) need to be detected the same as Mac OS X in terms of nan.h et al.  I'm pretty sure that that arrangement of headers as well as the lack of an fmodl is common to all systems with the BSD LIBC.
>>>
>>>Owen
>>>
>>>
>>>
>>