Re: [Plib-users] Failure in using plib ...

[Steve B. <sjb...@ai...>]

Digvijay Singh Lamba wrote:
> 
> Hi!
> I have recently compiled and installed plib on my redhat 7.0 system.

Aaaaarrrrrgggghhhhhhh!!!!!!

(Sorry - I said I'd scream if I heard another RH 7.0 problem).

RH7 is a 'lemon'.  They completely screwed up the compiler version
they shipped, the kernel header files that get installed by default
don't match the kernel that's installed...there are NUMEROUS other
screwups.

Personally, I'd be dumping RH7 and looking at some other distro (and
trying to get my money back from RH)...I use SuSE 7.0 and it seems fine.

As to what you should do...this is what I've heard (second hand) from
other mailing lists where a variety of strange problems are cropping
up:

The major beef is probably the C/C++ compiler.  The Bozo's at RH picked
an UNRELEASED CVS version of gcc - so who knows what problems it might
have?   Certainly other projects have seen all kinds of problems and
the Linux Kernel team created such a stink over RH using that half-assed
version that they demanded that RH install an older version of gcc for
rebuilding the kernel!  (That means that RH *knew* there was a problem
before they cut their CD-ROMS - and *STILL* went ahead and did it!)

So, the advice I've heard from other RH7 victims is to remove gcc/g++
and instead to symbolic-link the kernel's private compiler "kgcc"
(which is a nice stable version) to gcc.

At any rate, the check in ssgInit (and others) are simply:

static bool glIsValidContext ()
{
  return ( glXGetCurrentContext() != NULL ) ;
}

void ssgInit ()
{
  if ( ! glIsValidContext () )
  {
    ulSetError ( UL_FATAL, "ssgInit called without a valid OpenGL context.");
  }

  ...
}


So, there is no doubt that there is no bug in there.  AFAIK, there
is no problem with the released example programs.


Personally, I think every RH7 owner should demand their money back.


> this happens even with the standard examples with the lib package. Iam
> sure iam doing all the opengl initialisation correctly. Iam clueless what
> to do.
> I tried checking for the context immediately before the call to ssgInit
> and the context is fine at that place.

So what happened between YOUR call to glXGetCurrentContext() and mine?

The answer is "nothing" - so we are looking at a compiler problem - which
brings us back to RH7.

> also the examples for the sl and sg libraries run fine on my system.

Neither of those check OpenGL rendering contexts.

> PS: there was also another error, in the file that the automatic
> compilation created the js.h file did not include string.h for linux and
> so memcpy function was not bieng recognised.... had to do it manually,
> hope this is corrected in future.

Yes - sorry - this is a known problem...it'll get fixed in the next release.

All the programs I was testing with had:

   #include <string.h>
   #include <plib/js.h>

...so it wasn't showing up the problem...sorry!


-- 
Steve Baker   HomeEmail: <sjb...@ai...>
              WorkEmail: <sj...@li...>
              HomePage : http://web2.airmail.net/sjbaker1
              Projects : http://plib.sourceforge.net
                         http://tuxaqfh.sourceforge.net
                         http://tuxkart.sourceforge.net
                         http://prettypoly.sourceforge.net