Re: [openvrml-develop] openvrml-0.16.5: strange behaviour on FreeBSD.

[Braden M. <br...@en...>]

On Sun, 2007-06-03 at 22:29 +0200, Thierry Thomas wrote:
> Hello,
> 
> Trying to upgrade the FreeBSD port - still v. 0.15.10 - I have noticed
> some points:
> 
> - it builds fine, but I have to link openvrml-xembed and the tests
> programs with JNI_LIBS, ${JAVA_HOME}/jre/lib/${ARCH}/client/libjvm.so;

You must have enabled the Script node Java support. That isn't advised.
This feature isn't really supported in the current release. It doesn't
work well enough to be called "functional"; so turning it on definitely
isn't worth any problems doing so may cause.

> - to run it, I have to manually install libgdkglext-x11-1.0.so and
> libgtkglext-x11-1.0.so;

Weird. Perhaps this is something about the way libtool works (or
doesn't) on FreeBSD? gtkglext is included in the openvrml distribution
and the libraries you mention are used as "convenience libraries"--which
means that they should effectively be statically linked into
openvrml-xembed. That, anyway, is how it works on Linux.

> - openvrml-player can display e.g. the example Anchor.wrl, but trying
> warbird.wrl only displays the following messages:
> 
> ROLL_TURN: 3: Error: axis component of rotation is not normalized
> openvrml/script.cpp:3257: verification (ok) failed
> UP_DOWN_TURN: 5: Error: axis component of rotation is not normalized
> openvrml/script.cpp:3257: verification (ok) failed
> 35: Error: axis component of rotation is not normalized
> openvrml/script.cpp:3257: verification (ok) failed
> 35: Error: axis component of rotation is not normalized
> openvrml/script.cpp:3257: verification (ok) failed

These are normal. The warnings are benign; warbird.wrl just needs to be
cleaned up a bit to quell these.

> Assertion failed: (value == value), function x, file ./openvrml/basetypes.h, line 475.
> 
> (openvrml-player:81595): GLib-CRITICAL **: g_error_free: assertion `error != NULL' failed
> 
> Do you see these messages on other platforms?

This I haven't seen. (Or rather, I've seen things like it before, but
I'm not seeing it when I run warbird.wrl now.) Could you provide a stack
trace? (Best would be to open a new bug on SourceForge and attach it to
that.)

-- 
Braden McDaniel                           e-mail: <br...@en...>
<http://endoframe.com>                    Jabber: <br...@ja...>