I've uploaded 3.50c.  The binary is pretty good (fixed various performance
issues, now building with gcc 3.2), and I/O restrictions are enabled in
the binary but disabled in a usefully commented default
/usr/local/etc/povray.conf that is installed.  Note that the install
script won't clobber a pre-existing povray.conf, so you might want to
clear yours out of the way if you have one.  This version includes a
number of other bugfixes that caused build problems on Solaris and
BSD-derived Unices.

Minor caveat: The binary is built with Pentium optimizations.  It should
run fine (and that much faster) on Pentium or newer hardware.  If you're
running on less than a Pentium, you may need to recompile, and you may
want to consider a different hobby than ray-tracing. ;-)  If you have
sufficiently newer hardware and a sufficiently good compiler, you may want
to recompile for your specific arch, but that's up to you.

Major caveat: I realized after I'd finished uploading that the default
compiler flags also assume Pentium and are probably fairly gcc-specific. I
need to implement some decent configure options in that regard. Non-x86,
non-gcc users will need to edit src/Makefile for now.  Consider yourselves
warned.  I should have a fix up soon. The 3.50d binary probably won't be
different, unless there are other problems.

I did some research and found that I should distribute the object files I
got from the build if I want to distribute binaries which are statically
linked against LGPL libraries.  There's now a povobj.tar.gz file in the
ftp directory in case anyone wants them, along with a file containing the
link line I used.

-Mark Gordon

Post a reply to this message

Mark Gordon wrote:
> 
> I've uploaded 3.50c.  The binary is pretty good (fixed various performance
> issues, now building with gcc 3.2), and I/O restrictions are enabled in
> the binary but disabled in a usefully commented default
> /usr/local/etc/povray.conf that is installed.  Note that the install
> script won't clobber a pre-existing povray.conf, so you might want to
> clear yours out of the way if you have one.  This version includes a
> number of other bugfixes that caused build problems on Solaris and
> BSD-derived Unices.
> [...]

Nice, but i miss the list of changes from the first release to 3.50c (in
3.1 this was in 'revision.doc').  The 'ChangeLog' file's latest entry is
from 8 July 2002.

Christoph

-- 
POV-Ray tutorials, IsoWood include,                 
TransSkin and more: http://www.tu-bs.de/~y0013390/  
Last updated 13 Aug. 2002 _____./\/^>_*_<^\/\.______

Post a reply to this message

> Major caveat: I realized after I'd finished uploading that the default
> compiler flags also assume Pentium and are probably fairly gcc-specific. I
> need to implement some decent configure options in that regard.

	Here is an autoconf macro that you can use for this purpose
(put it in acinclude.m4):


# POV_TRY_CFLAGS(CFLAGS, [ACTION-IF-WORKS], [ACTION-IF-FAILS])
# ------------------------------------------------------------
# Check whether compiler supports a given set of cflags, cache result,
# and update CFLAGS.
#
AC_DEFUN([POV_TRY_CFLAGS],
[
   # Create a unique cache-id name.
   pov_try_cflags_var=pov_cv_try_cflags`echo "$1" | sed 's/[[^a-zA-Z0-9]]/_/g'`
   # Cannot use AC_CACHE_CHECK() due to the nature of the cache-id variable.
   AC_MSG_CHECKING([whether $CC accepts $1])
   AC_CACHE_VAL(
     [$pov_try_cflags_var],
     [
       # AC_TRY_COMPILE() is not appropriate any longer since autoconf-2.5x.
       # Compile and inspect standard error for given flags.
       echo 'int main(void){ return 0; }' > conftest.c
       if test -z "`$CC -c $1 conftest.c 2>&1 | grep '\$1'`"; then
         eval "$pov_try_cflags_var=yes"
       else
         eval "$pov_try_cflags_var=no"
       fi
       rm -f conftest.c conftest.o
     ]
   )
   eval "pov_try_cflags_value=\$$pov_try_cflags_var"
   AC_MSG_RESULT([$pov_try_cflags_value])

   # Update CFLAGS when flags are working, and run provided actions.
   if test x"$pov_try_cflags_value" = x"yes"; then
     CFLAGS="$CFLAGS $1"
     ifelse([$2],[],[:],[$2])
   else
     ifelse([$3],[],[:],[$3])
   fi
])


	In the configure.ac (or configure.in) file you just call it,
for instance:

TRIP_TRY_CFLAGS([-O2], [], [TRIP_TRY_CFLAGS([-O])])

	will try first flag -O2 and, if not recognized, try -O and then
simply nothing. CFLAGS is updated accordingly and the result is cached.

	This macro normally works okay with gcc on Linux and gcc/cc on
SGI (didn't tested anything else). Hope I didn't put any typo while
copying/editing the macro from my own packages devels.

	- N.C.

Post a reply to this message

> TRIP_TRY_CFLAGS([-O2], [], [TRIP_TRY_CFLAGS([-O])])

	Sorry this is obviously:

POV_TRY_CFLAGS([-O2], [], [POV_TRY_CFLAGS([-O])])

	By the way this macro is for usage with autoconf-2.5x.
I didn't tested it with the old autoconf-2.13 which might give
problem with the quotations used in the sed call of the macro
for instance. All arguments must be properly quoted with brackets,
otherwise the macro will fail.

	- N.C.


PS: some other examples:

POV_TRY_CFLAGS([-ansi], [], [POV_TRY_CFLAGS([-Aa])])
POV_TRY_CFLAGS([-Wall], [], [POV_TRY_CFLAGS([-fullwarn])])

Post a reply to this message

Thanks Gordon, great.

The spec file is not yet included. Is there any serious reason
for this, like 'doesn't work properly'? Or just lack of time ;-/  ?

-Hans-

Mark Gordon wrote:
> I've uploaded 3.50c.  The binary is pretty good (fixed various performance
> issues, now building with gcc 3.2), and I/O restrictions are enabled in
> the binary but disabled in a usefully commented default
> /usr/local/etc/povray.conf that is installed.  Note that the install
> script won't clobber a pre-existing povray.conf, so you might want to
> clear yours out of the way if you have one.  This version includes a
> number of other bugfixes that caused build problems on Solaris and
> BSD-derived Unices.
> 
> Minor caveat: The binary is built with Pentium optimizations.  It should
> run fine (and that much faster) on Pentium or newer hardware.  If you're
> running on less than a Pentium, you may need to recompile, and you may
> want to consider a different hobby than ray-tracing. ;-)  If you have
> sufficiently newer hardware and a sufficiently good compiler, you may want
> to recompile for your specific arch, but that's up to you.
> 
> Major caveat: I realized after I'd finished uploading that the default
> compiler flags also assume Pentium and are probably fairly gcc-specific. I
> need to implement some decent configure options in that regard. Non-x86,
> non-gcc users will need to edit src/Makefile for now.  Consider yourselves
> warned.  I should have a fix up soon. The 3.50d binary probably won't be
> different, unless there are other problems.
> 
> I did some research and found that I should distribute the object files I
> got from the build if I want to distribute binaries which are statically
> linked against LGPL libraries.  There's now a povobj.tar.gz file in the
> ftp directory in case anyone wants them, along with a file containing the
> link line I used.
> 
> -Mark Gordon

Post a reply to this message

Hans-Detlev Fink wrote:
> Thanks Gordon, great.
> 

Thanks *Mark* of course! Sorry for that :)

Post a reply to this message

On Wed, 23 Oct 2002 14:00:20 +0200, Hans-Detlev Fink wrote:

> Hans-Detlev Fink wrote:
>> Thanks Gordon, great.
>> 
>> 
> Thanks *Mark* of course! Sorry for that :)

I hadn't heard back before I made the last package.  It will be in 3.50d.

-Mark Gordon

Post a reply to this message

In article <pan### [at] povrayorg> , "Mark Gordon" 
<mtg### [at] povrayorg> wrote:

> I've uploaded 3.50c.

BTW, is the Unix source the same as the Linux source?  I am wondering
because the source archive in the Unix dir has a different date...

    Thorsten

____________________________________________________
Thorsten Froehlich, Duisburg, Germany
e-mail: tho### [at] trfde

Visit POV-Ray on the web: http://mac.povray.org

Post a reply to this message

On Wed, 23 Oct 2002 08:06:26 +0000, Mark Gordon wrote:
> I hadn't heard back before I made the last package.  It will be in 3.50d.
> 
> -Mark Gordon

I went ahead and built a Pentium-III optimized version as an RPM
using GCC 3.2 on Red Hat 8.0.  I used Dan's spec file as the starting
point.  There's a source RPM as well.

See:  http://www.integrated-dynamics.com/gar for information.

Feedback and suggestions welcome, as well.

Gordon.
-- 
"Far and away the best prize that life has to offer
  is the chance to work hard at work worth doing."
       -- Theodore Roosevelt

Post a reply to this message

On Wed, 23 Oct 2002 16:31:21 +0200, Thorsten Froehlich wrote:

> In article <pan### [at] povrayorg> , "Mark Gordon" 
> <mtg### [at] povrayorg> wrote:
> 
>> I've uploaded 3.50c.
> 
> BTW, is the Unix source the same as the Linux source?  I am wondering
> because the source archive in the Unix dir has a different date...
> 
>     Thorsten

The source tarball in the Unix directory is a symlink to the source
tarball in the Linux directory.  I believe the date reflects when the
symlink was created.

-Mark Gordon

Post a reply to this message