[osg-submissions] osgPluging pfb -- Mutlitexturing Supported & Segmentation fault fixed

Wed Jul 16 05:14:35 PDT 2008

Hi Rober,

This can be achieved easily.
pf.h defines PF_MAJOR_VERSION, PF_MINOR_VERSION & PF_MAINT_VERSION.
I guess all is left is the polling to the users and/or some testing from
someone with older version than me in case this is not needed.

Anthousis

On Wed, Jul 16, 2008 at 11:42 AM, Robert Osfield <robert.osfield at gmail.com>
wrote:

> Hi Anthousis,
>
> I think what we may have to do is poll the osg-users for feedback on
> the multi-texture API in Performer, and suspect perhaps the best way
> to tackle it is to do an #ifdef PERFORMER_VERSION to retain the old
> path that is the older single texture API whilst providing the new
> multi-texture path that you've developed.  If we can't auto-detect the
> Performer version then we'd need to put it in as an option via CMake.
>
> Robert.
>
> On Wed, Jul 16, 2008 at 11:21 AM, Anthousis Andreadis
> <anthousis at gmail.com> wrote:
> > Hi Robert,
> >
> > 1) I am using v3.2.1 of Performer
> > 3) I do is the following.
> >
> >    For each unit i check whether data exists (texture in our case)
> >          if true
> >             copy that to the osg data structure
> >          else
> >             continue
> >
> >    (the file i attach has for (tex_unit=0; tex_unit<PF_MAX_TEXTURES
> instead
> > of 8 i had in the previous. Sorry for that again)
> >
> >    In that way when a pfGeoSet has only one texture the corresponding
> > StateSet will contain only that texture.
> >    If someone on the other hand has stored 6 textures and 6 texture
> > coordinate sets (for shader use ofcourse) all data will
> >    be passed to the osg data structure as well.
> >
> >    I use both non multitextured models and multitextured.
> >
> >   The only problem is i don't have a fast answer for your 2nd question
> since
> > this is the first version of performer i ever used. I will try to find
> that
> > out.
> >
> > Anthousis
> >
> > ps. Sorry again for the re-attaching of file. The previous didn't cause
> any
> > problem but this is more correct.
> >
> > On Wed, Jul 16, 2008 at 9:42 AM, Robert Osfield <
> robert.osfield at gmail.com>
> > wrote:
> >>
> >> Hi Anthousis,
> >>
> >> Thanks for the smaller set of changes, it's now possible for me to do
> >> a proper review.
> >>
> >> A couple of questions, to help clarify things.
> >>
> >>  1) Which version of Performer are you compiling against?
> >>  2) Which Performer versions support the multi-texture interface?
> >>  3) What happens what a pfGeoSet doesn't use the multi-texture
> interface?
> >>
> >> My concerns is that we might break the build for other users not using
> >> a modern Performer version, and that the older single texture and more
> >> modern multi-texture interface don't overlap i.e. what happens when a
> >> model has no multi-textures on it.
> >>
> >> Robert.
> >>
> >>
> >> On Wed, Jul 16, 2008 at 10:09 AM, Anthousis Andreadis
> >> <anthousis at gmail.com> wrote:
> >> > Hi Robert,
> >> >
> >> > Sorry for that. I used an indent script i use on my source (i forgot
> it)
> >> > and
> >> > you saw the results...
> >> >
> >> > I also checked both the cases i mentioned with svn version and i
> >> > couldn't
> >> > reproduce the segmentation fault which i originally noticed in 2.2.0
> >> > version. That's the reason i attach only the files for the
> >> > multitexturing
> >> > support of pfb files.
> >> >
> >> > This time they contain only the changes.
> >> >
> >> > Anthousis
> >> >
> >> > On Tue, Jul 15, 2008 at 6:37 PM, Robert Osfield
> >> > <robert.osfield at gmail.com>
> >> > wrote:
> >> >>
> >> >> Hi Anthousus,
> >> >>
> >> >> I've just done a graphical diff between your changed files and the
> >> >> ones in SVN trunk, and there are a massive number of changes, most of
> >> >> which seem to surperflous indention/coding style changes.  The extent
> >> >> of these changes means that it very very hard to spot the actual
> >> >> important changes that you have made.  It's just too much of a risk
> to
> >> >> merge a whole bag of other changes when they don't actual address a
> >> >> problem - the risk is too high on breaking other bits of the code.
> >> >>
> >> >> Could you please explain why you took on yourself to go change so
> many
> >> >> things, and what bits are the important functional bits of your
> >> >> changes.  Could you also please use the SVN as base for your changes,
> >> >> and apply only changes your need for the functionality you need, once
> >> >> I have this set of changes I can undertake a proper review.
> >> >>
> >> >> Thanks,
> >> >> Robert.
> >> >>
> >> >> On Tue, Jul 15, 2008 at 1:09 PM, Anthousis Andreadis
> >> >> <anthousis at gmail.com> wrote:
> >> >> > Hi
> >> >> >
> >> >> > The 2 problems : osgPlugin::pfb
> >> >> >
> >> >> > 1)
> >> >> > When loading pfb files only the first texture is loaded.
> >> >> > Fixed in-order to check all available texture positions.
> >> >> >
> >> >> >
> >> >> > Result:: support for multitextured pfb files
> >> >> >
> >> >> > 2)
> >> >> > When loading the same .pfb file twice you get a segmentation fault.
> >> >> > I fixed this by replacing the "unrefDelete" command with the
> >> >> > performer
> >> >> > command "pfDelete"
> >> >> >
> >> >> >
> >> >> >
> >> >> > The changes have been tested with many pfb models both
> multitextured
> >> >> > and
> >> >> > not.
> >> >> >
> >> >> > Anthousis Andreadis
> >> >> >
> >> >> > _______________________________________________
> >> >> > osg-submissions mailing list
> >> >> > osg-submissions at lists.openscenegraph.org
> >> >> >
> >> >> >
> >> >> >
> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
> >> >> >
> >> >> >
> >> >> _______________________________________________
> >> >> osg-submissions mailing list
> >> >> osg-submissions at lists.openscenegraph.org
> >> >>
> >> >>
> >> >>
> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
> >> >
> >> >
> >> > _______________________________________________
> >> > osg-submissions mailing list
> >> > osg-submissions at lists.openscenegraph.org
> >> >
> >> >
> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
> >> >
> >> >
> >> _______________________________________________
> >> osg-submissions mailing list
> >> osg-submissions at lists.openscenegraph.org
> >>
> >>
> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
> >
> >
> > _______________________________________________
> > osg-submissions mailing list
> > osg-submissions at lists.openscenegraph.org
> >
> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
> >
> >
> _______________________________________________
> osg-submissions mailing list
> osg-submissions at lists.openscenegraph.org
>
> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openscenegraph.org/pipermail/osg-submissions-openscenegraph.org/attachments/20080716/8c05e832/attachment-0001.htm>