[info-performer] Performer's flt db loader

New Message Reply Date view Thread view Subject view Author view

From: Rob Krajcarski (rjkrajca++at++dciem.dnd.ca)
Date: 04/18/2002 02:01:17


Hello,

I'm in the middle of trying to move (or at least test) an application we
have running on SGIs over to Linux. First of all, let me warn you that
our SGIs are currently running Performer 2.2.11, while the Linux box is
running 2.5.0-5. So, no suprise, I've run into a few problems during this
transition. Currently I'm working on problems that seem to be as a result
of changes to the pfflt database loader. Now, I'm not sure if the
problems are as a result of OS (sgi->linux), Performer version
(2.2.11->2.5.0), or some weird combination of the two. The specific
problems that I'm dealing with are:

1) Bead id's seem to get corrupted. I've found that bead id's whose
length is of a specific length (I've seen strlen of 8, 12 and 16) cause
the pfNode name to be followed by a "\n", causing all pfNode->find() calls
to fail miserably. I'm pretty sure I remember seeing something about
problems when the length of the bead id is a multiple of 8, but I can't
remember where I saw this, or the details (and the string of length 12
just messes all that up!).

2) Selected elements of the graph seem to be excluded when loading the
file. I have a program that loads a file in (after disabling cleaning and
flattening) and simply prints out the contents. On the SGI everything is
fine, unlike Linux. The graph I'm working with is pretty big, so I
haven't checked out all the differences, but what I did notice was that I
had some light sources that were excluded. The graph includes one group
node with 11 children, 3 of which are light sources. Well, under linux 2
of the 3 light sources just don't show up.

Anyway, I hope I haven't been too long winded (I'm in the middle of
'documentation' stage of development, so I seem to be getting more wordy
:)

Any suggestions, or pointers to were I could find out what could be
causing these problems would be greatly appreciated.

Rob Krajcarski
DRDC-Toronto


New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Apr 18 2002 - 07:09:35 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.