Re: [osg-users] VPB in SVN

2008-01-09 Thread Robert Osfield
Hi J-S,

VPB from SVN does need porting to Windows, in particular the posix
functions used for managing the file systems need to have equivalents
found.  I'd appreciate Windows users tackling this as I don't have the
Windows here to test and work on.

A sort cut to Windows support would also be to compile using Cygwin as
this will provide all the posix functions for you.

W.r.t warnings, g++ 4.2.1 isn't reporting any warnings, I haven't
upped the warning level from default so perhaps this might winkle so
more out.  I guess some of the warnings might be related to just
targeting posix so far.

W.r.t fallback that does work under Windows, 0.9.1 should work fine.

Robert.

On Jan 9, 2008 3:00 PM, Jean-Sébastien Guay
<[EMAIL PROTECTED]> wrote:
> Hello Robert,
>
> I tried fetching VPB from SVN and compiling it on Windows, and I
> wonder if the current SVN version is in a state of flux. Apart from a
> few errors that are down to it not having been tested on Windows,
> there are a lot of errors that should be flagged on any compiler
> (methods not returning a value when they should, functions not
> included, etc.).
>
> We can use 0.9.1 for the time being, but is this expected?
>
> Thanks,
>
> J-S
> --
> __
> Jean-Sebastien Guay [EMAIL PROTECTED]
>  http://whitestar02.webhop.org/
>
> 
> This message was sent using IMP, the Internet Messaging Program.
>
>
> ___
> osg-users mailing list
> osg-users@lists.openscenegraph.org
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-09 Thread Jean-Sébastien Guay
Hello Robert,

> VPB from SVN does need porting to Windows, in particular the posix
> functions used for managing the file systems need to have equivalents
> found.  I'd appreciate Windows users tackling this as I don't have the
> Windows here to test and work on.

OK, I'll probably be able to help there in the coming weeks.

> W.r.t warnings, g++ 4.2.1 isn't reporting any warnings, I haven't
> upped the warning level from default so perhaps this might winkle so
> more out.  I guess some of the warnings might be related to just
> targeting posix so far.

I didn't mean warnings, there are some errors that should show up on  
any compiler, such as a method declared as bool someMethod() not  
returning a value... For example, in include/vpb/FileCache :

 bool sync() { if (_requiresWrite) write(_filename); }

should be

 bool sync() { if (_requiresWrite) return write(_filename); return false; }

or something to that effect. I would have thought gcc would give an  
error there... MSVC certainly does.

BTW, to correct the methods, I would like to know what is your  
standard for methods returning error? 0/false = no error and 1/true =  
error? Or 0/false = failed and 1/true = succeeded?

> W.r.t fallback that does work under Windows, 0.9.1 should work fine.

Yep, works fine.

Thanks,

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-09 Thread David Callu
Hi Robert,


Have you put "-W -Wall" in CMAKE_CXX_FLAGS.
I have a lot of warning too, but I know that your code work fine with
this warning.
So I never take the time to fix them.

Cheers
David

2008/1/9, Robert Osfield <[EMAIL PROTECTED]>:
>
> On Jan 9, 2008 3:45 PM, Jean-Sébastien Guay
> <[EMAIL PROTECTED]> wrote:
> > OK, I'll probably be able to help there in the coming weeks.
>
> Great, much appreciated.
>
> > I didn't mean warnings, there are some errors that should show up on
> > any compiler, such as a method declared as bool someMethod() not
> > returning a value... For example, in include/vpb/FileCache :
> >
> >  bool sync() { if (_requiresWrite) write(_filename); }
> >
> > should be
> >
> >  bool sync() { if (_requiresWrite) return write(_filename); return
> false; }
> >
> > or something to that effect. I would have thought gcc would give an
> > error there... MSVC certainly does.
>
> I've just done a clean build and g++ isn't reporting any warnings
> whatsoever.
>
> The above is an error of course, I'll fix this.
>
>
> > BTW, to correct the methods, I would like to know what is your
> > standard for methods returning error? 0/false = no error and 1/true =
> > error? Or 0/false = failed and 1/true = succeeded?
>
> Typically true for succeeded/did operation.
>
> Robert.
> ___
> osg-users mailing list
> osg-users@lists.openscenegraph.org
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-09 Thread Robert Osfield
On Jan 9, 2008 3:45 PM, Jean-Sébastien Guay
<[EMAIL PROTECTED]> wrote:
> OK, I'll probably be able to help there in the coming weeks.

Great, much appreciated.

> I didn't mean warnings, there are some errors that should show up on
> any compiler, such as a method declared as bool someMethod() not
> returning a value... For example, in include/vpb/FileCache :
>
>  bool sync() { if (_requiresWrite) write(_filename); }
>
> should be
>
>  bool sync() { if (_requiresWrite) return write(_filename); return false; 
> }
>
> or something to that effect. I would have thought gcc would give an
> error there... MSVC certainly does.

I've just done a clean build and g++ isn't reporting any warnings whatsoever.

The above is an error of course, I'll fix this.


> BTW, to correct the methods, I would like to know what is your
> standard for methods returning error? 0/false = no error and 1/true =
> error? Or 0/false = failed and 1/true = succeeded?

Typically true for succeeded/did operation.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Christophe Lombardo
Hi

I have a  msvc8 version of vpb 0.9.3.
I found some straight forward equivalents for most of the posix 
functions, but I had to make rather rude stuff for others.
I did not test this part (tasks and machines), but the project compiles 
and run fine on one single computer.

I don't know whether this can be useful or not, nor how I can 
contribute. Let me know.

jcl

-- 
Jean-Christophe Lombardo  - Salle Immersive "Le Corbusier"
http://salle-immersive.cstb.fr
EVE / Centre Scientifique et Technique du Batiment
290 route des Lucioles - BP 209 - 06904 Sophia Antipolis - France 




 
 

This footnote confirms that this email message has been scanned by
PineApp Mail-SeCure for the presence of malicious code, vandals & computer 
viruses.




___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Robert Osfield
Hi JCL,

I am just merging changes than JS submitted yesterday evening, I'm not
merging all the suggested changes, rather merging the uncontroversial
ones first, making other alternative ammendments for others.  These
changes are now checked in.

A number of items are outstanding, such as the Windows versions of
posix functions.  JS has created a Utils file for these, I'm about to
just start reviewing these.  I'll follow JS's lead in collecting all
these functions into one place, but I'll probably not follow the
naming convention he's adopted.   I will make further steps along
getting route to getting things compiling under Windows.

My suggestion for yourself is to wait for me to get the next set of
changes merged and submitted.  This probably won't get things working
out of the box under Windows quite yet, but should make it clearer
what bits to tweak and where to complete the job.  At this point it'd
be very useful if you could help test and fill out the blanks where
they are missing.

Robert.

On Jan 10, 2008 10:43 AM, Jean-Christophe Lombardo
<[EMAIL PROTECTED]> wrote:
> Hi
>
> I have a  msvc8 version of vpb 0.9.3.
> I found some straight forward equivalents for most of the posix
> functions, but I had to make rather rude stuff for others.
> I did not test this part (tasks and machines), but the project compiles
> and run fine on one single computer.
>
> I don't know whether this can be useful or not, nor how I can
> contribute. Let me know.
>
> jcl
>
> --
> Jean-Christophe Lombardo  - Salle Immersive "Le Corbusier"
> http://salle-immersive.cstb.fr
> EVE / Centre Scientifique et Technique du Batiment
> 290 route des Lucioles - BP 209 - 06904 Sophia Antipolis - France
>
>
>
>
>
>
> 
> This footnote confirms that this email message has been scanned by
> PineApp Mail-SeCure for the presence of malicious code, vandals & computer 
> viruses.
> 
>
>
>
>
> ___
> osg-users mailing list
> osg-users@lists.openscenegraph.org
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Christophe Lombardo
ok, fine for me.

jcl



 
 

This footnote confirms that this email message has been scanned by
PineApp Mail-SeCure for the presence of malicious code, vandals & computer 
viruses.




___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Robert Osfield
On Jan 10, 2008 11:13 AM, Jean-Christophe Lombardo
<[EMAIL PROTECTED]> wrote:
> ok, fine for me.

The changes are now all checked in.  Could you do an svn update on OSG
and VPB and let me know how you get on.

Also could you review the src/vpb/FileUtils.cpp file to see if the
windows equivalents of posix methods match up to what you tried, if
you spot anything that could be improved upon just email the osg-users
list with the suggestions.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Sébastien Guay
Bonjour Jean-Christophe,

> I don't know whether this can be useful or not, nor how I can
> contribute. Let me know.

Hehe, I would have liked to know this yesterday :-)

I sent to osg-submissions the changed files to make VPB from SVN  
compile cleanly on Windows yesterday. We'll see if they make it into  
SVN as-is or if more work is needed. And similarly to you, I have not  
yet tested the parts that use posix functions, just good old osgdem...

And you're right, most functions have straight equivalents (or just  
add an underscore to the name), but others were a bit more touchy. In  
any case, these things rarely work right after a straight port, as  
platform compatibility often amounts to more than just function names.

Hopefully it's a step in the right direction. Thanks for your offer,  
even if it was just about 12 hours late :-)

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Sébastien Guay
Hi Robert,

> A number of items are outstanding, such as the Windows versions of
> posix functions.  JS has created a Utils file for these, I'm about to
> just start reviewing these.  I'll follow JS's lead in collecting all
> these functions into one place, but I'll probably not follow the
> naming convention he's adopted.   I will make further steps along
> getting route to getting things compiling under Windows.

Changing the naming convention is fine. I just picked one that seemed  
ok, in the hopes of avoiding peppering the whole code-base with #ifdef  
WIN32 #else #endif everywhere...

Thanks,

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Christophe Lombardo
I still have 2 slight modifications to suggest on windows side:
 * in vpb/FileUtils I had to add a #include   to define the 
O_RDWR flag with no leading _
 * in vpb/FileUtils.cpp I think that vpb::sync() could be defined as 
void vpb::sync() { (void) _flushall(); }

Jean-Sébastien, I'm sorry for the 12 hours lag, but 12 hours ago, I was 
sleeping :) as it was night on my side of the pond...

jcl




 
 

This footnote confirms that this email message has been scanned by
PineApp Mail-SeCure for the presence of malicious code, vandals & computer 
viruses.




___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Sébastien Guay
Hello Robert,

> Fingers crossed this should now mean VPB is
> ported back to windows, at least for basic osgdem functionality.

Yep, seems so. osgdem still running... I probably should have chosen a  
smaller dataset as a first test. :-)

> Once I've completed some more work on
> VPB I'll do a write up on all the new functionality and then strike up
> a discussion on osg-users about how to tackle distributed builds under
> Windows/mixed networks, I'll need suggestions from the community on
> this part though, the best I can do is say how things work under unix
> and trust in windows experts for a similar solution under Windows.

Looking forward to it. I imagine if distributed builds work, it should  
be possible to use multiprocessor/multicore machines in the same way,  
by just specifying the same machine multiple times? Right now osgdem  
is taking only 25% of my spiffy quad-core CPU! :-(

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Sébastien Guay
Hello jcl,

>  * in vpb/FileUtils I had to add a #include   to define the
> O_RDWR flag with no leading _

Already done in the second batch of modifications (it was already in  
the #else, so we just moved it out of the #ifdef).

>  * in vpb/FileUtils.cpp I think that vpb::sync() could be defined as
> void vpb::sync() { (void) _flushall(); }

Nice, I didn't know about that one. I did some research on an  
equivalent to sync() on Windows, and that one never came up... Could  
you add it please Robert?

> Jean-Sébastien, I'm sorry for the 12 hours lag, but 12 hours ago, I was
> sleeping :) as it was night on my side of the pond...

I know, I know :-) I was just joking. That's the interesting thing  
about contributing to this project, during the afternoon there is  
almost no traffic on the list, and then when I get back in the morning  
there's a slew of new posts - Robert being in Scotland and all... :-)

Thanks for testing.

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Robert Osfield
On Jan 10, 2008 1:45 PM, Robert Osfield <[EMAIL PROTECTED]> wrote:
> The changes are now all checked in.  Could you do an svn update on OSG
> and VPB and let me know how you get on.

I've just received, merged and submitted to SVN further update from
J-S that tweak the changes I checked in earlier, so another svn update
is required for VPB.  Fingers crossed this should now mean VPB is
ported back to windows, at least for basic osgdem functionality.

While VPB should now compile under Windows the distributed builds are
likely to still be a problem as they rely upon network files systems
and remote application invocation using ssh are all used, both things
that are possible under Windows, but not quite as straight
forward/built in as in unix.  Once I've completed some more work on
VPB I'll do a write up on all the new functionality and then strike up
a discussion on osg-users about how to tackle distributed builds under
Windows/mixed networks, I'll need suggestions from the community on
this part though, the best I can do is say how things work under unix
and trust in windows experts for a similar solution under Windows.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Robert Osfield
On Jan 10, 2008 3:16 PM, Jean-Sébastien Guay
<[EMAIL PROTECTED]> wrote:
> Yep, seems so. osgdem still running... I probably should have chosen a
> smaller dataset as a first test. :-)

If its a dataset with DEM's try adding --terrain to the command line,
it builds much, much faster using osgTerrain::Terrain to store the
GeightFields rather than creating osg::Geometry meshes from the DEM's
and then simplifying these (its the simplification which is slow).
The --terrain functionality is very early days though (checked in
yesterday) so there are still a few loose ends, but its certainly
enough to have a first pass look.

> > Once I've completed some more work on
> > VPB I'll do a write up on all the new functionality and then strike up
> > a discussion on osg-users about how to tackle distributed builds under
> > Windows/mixed networks, I'll need suggestions from the community on
> > this part though, the best I can do is say how things work under unix
> > and trust in windows experts for a similar solution under Windows.
>
> Looking forward to it. I imagine if distributed builds work, it should
> be possible to use multiprocessor/multicore machines in the same way,
> by just specifying the same machine multiple times? Right now osgdem
> is taking only 25% of my spiffy quad-core CPU! :-(

Yep, you can use multiple cores, and multiple machines, or multiple
cores on one machine, all managed via the new vpbmaster app.  osgdem
itself has some preliminary multi-threaded support to, but not yet to
the level afforded by running multiple osgdem's in parallel.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Paul Martz
I just wanted to thank all involved for the work on this.

Bob and I wanted to use SVN head of VPB for the upcoming terrain training
January 24, but aborted that idea and went back to 0.9.1 when we encountered
the Windows issues. I was reluctant to bite off porting it with so much to
do between now and the course.

With this work in place, Bob and I will once again consider using SVN head
for the course.

(Shameless plug: register for the course here:
http://www.skew-matrix.com/training.asp.)

Thanks again,

Paul Martz
Skew Matrix Software LLC
http://www.skew-matrix.com
303 859 9466

___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Robert Osfield
On Jan 10, 2008 3:15 PM, Jean-Christophe Lombardo
<[EMAIL PROTECTED]> wrote:
> I still have 2 slight modifications to suggest on windows side:
>  * in vpb/FileUtils I had to add a #include   to define the
> O_RDWR flag with no leading _

Already checked in.

>  * in vpb/FileUtils.cpp I think that vpb::sync() could be defined as
> void vpb::sync() { (void) _flushall(); }

Thanks, I've added this and checked it in.

> Jean-Sébastien, I'm sorry for the 12 hours lag, but 12 hours ago, I was
> sleeping :) as it was night on my side of the pond...

OSG/VPB development never sleeps :-)

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Sébastien Guay
Hi Robert,

> The --terrain functionality is very early days though (checked in
> yesterday) so there are still a few loose ends, but its certainly
> enough to have a first pass look.

Cool, I'll try that out.

> Yep, you can use multiple cores, and multiple machines, or multiple
> cores on one machine, all managed via the new vpbmaster app.  osgdem
> itself has some preliminary multi-threaded support to, but not yet to
> the level afforded by running multiple osgdem's in parallel.

Excellent!

Thanks,

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Sébastien Guay
Hi Paul,

> Bob and I wanted to use SVN head of VPB for the upcoming terrain training
> January 24, but aborted that idea and went back to 0.9.1 when we encountered
> the Windows issues. I was reluctant to bite off porting it with so much to
> do between now and the course.

That was actually my motivation too. :-) When I saw that you were  
recommending downloading 0.9.1, I got the SVN, tried it, and saw it  
didn't compile...

Looking forward to the course.

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Sébastien Guay
Hello Robert,

> If its a dataset with DEM's try adding --terrain to the command line

If that gives me a file which doesn't display anything (when given to  
osgviewer) does that mean that --terrain didn't work, or does it mean  
that the dataset didn't have DEMs? I *think* I'm using geotiffs (still  
very new to this!).

Is there any reference to the various formats that osgdem can use? I  
could wait until Paul and Bob's course, but I'm curious :-)

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Sébastien Guay
Hello Robert,

> --terrain should work fine with imagery only databases, not seeing
> something suggests problems somewhere, but where? How knows with so
> little info to go on.  Does the generated database run without if you
> don't use the --terrain option?

Sorry for the little info, don't really know what to give. The osgdem  
without --terrain just finished, it worked and looks great. Running  
with --terrain ran a lot faster, but I can't see anything in osgviewer  
with the resulting file. The file is over twice the size (~500k versus  
~202k without --terrain). It doesn't really matter either way - I was  
just trying it out.

> gdal is used for loaded all imagery and dem's so simpy run
>
>gdalinfo --formats

Thanks, I'll look those up. Geotiffs are the most common/useful, is  
that correct? (just assuming from what I've read in the past on this  
list)

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Robert Osfield
On Jan 10, 2008 4:47 PM, Jean-Sébastien Guay
<[EMAIL PROTECTED]> wrote:
> Hello Robert,
>
> > If its a dataset with DEM's try adding --terrain to the command line
>
> If that gives me a file which doesn't display anything (when given to
> osgviewer) does that mean that --terrain didn't work, or does it mean
> that the dataset didn't have DEMs? I *think* I'm using geotiffs (still
> very new to this!).

--terrain should work fine with imagery only databases, not seeing
something suggests problems somewhere, but where? How knows with so
little info to go on.  Does the generated database run without if you
don't use the --terrain option?

> Is there any reference to the various formats that osgdem can use? I
> could wait until Paul and Bob's course, but I'm curious :-)

gdal is used for loaded all imagery and dem's so simpy run

   gdalinfo --formats

When I do this on my machine I get:

Supported Formats:
  VRT (rw+): Virtual Raster
  GTiff (rw+): GeoTIFF
  NITF (rw+): National Imagery Transmission Format
  HFA (rw+): Erdas Imagine Images (.img)
  SAR_CEOS (ro): CEOS SAR Image
  CEOS (ro): CEOS Image
  ELAS (rw+): ELAS
  AIG (ro): Arc/Info Binary Grid
  AAIGrid (rw): Arc/Info ASCII Grid
  SDTS (ro): SDTS Raster
  OGDI (ro): OGDI Bridge
  DTED (rw): DTED Elevation Raster
  PNG (rw): Portable Network Graphics
  JPEG (rw): JPEG JFIF
  MEM (rw+): In Memory Raster
  JDEM (ro): Japanese DEM (.mem)
  GIF (rw): Graphics Interchange Format (.gif)
  ESAT (ro): Envisat Image Format
  BSB (ro): Maptech BSB Nautical Charts
  XPM (rw): X11 PixMap Format
  BMP (rw+): MS Windows Device Independent Bitmap
  AirSAR (ro): AirSAR Polarimetric Image
  RS2 (ro): RadarSat 2 XML Product
  PCIDSK (rw+): PCIDSK Database File
  PCRaster (rw): PCRaster Raster File
  ILWIS (rw+): ILWIS Raster Map
  SGI (ro): SGI Image File Format 1.0
  Leveller (ro): Leveller heightfield
  GMT (rw): GMT NetCDF Grid Format
  netCDF (rw): Network Common Data Format
  HDF4 (ro): Hierarchical Data Format Release 4
  HDF4Image (rw+): HDF4 Dataset
  PNM (rw+): Portable Pixmap Format (netpbm)
  DOQ1 (ro): USGS DOQ (Old Style)
  DOQ2 (ro): USGS DOQ (New Style)
  ENVI (rw+): ENVI .hdr Labelled
  EHdr (rw+): ESRI .hdr Labelled
  PAux (rw+): PCI .aux Labelled
  MFF (rw+): Vexcel MFF Raster
  MFF2 (rw+): Vexcel MFF2 (HKV) Raster
  FujiBAS (ro): Fuji BAS Scanner Image
  GSC (ro): GSC Geogrid
  FAST (ro): EOSAT FAST Format
  BT (rw+): VTP .bt (Binary Terrain) 1.3 Format
  LAN (ro): Erdas .LAN/.GIS
  CPG (ro): Convair PolGASP
  IDA (rw+): Image Data and Analysis
  NDF (ro): NLAPS Data Format
  DIPEx (ro): DIPEx
  ISIS2 (ro): USGS Astrogeology ISIS cube (Version 2)
  PDS (ro): NASA Planetary Data System
  JPEG2000 (rw): JPEG-2000 part 1 (ISO/IEC 15444-1)
  L1B (ro): NOAA Polar Orbiter Level 1b Data Set
  FIT (rw): FIT Image
  RMF (rw+): Raster Matrix Format
  WCS (ro): OGC Web Coverage Service
  RST (rw+): Idrisi Raster A.1
  RIK (ro): Swedish Grid RIK (.rik)
  USGSDEM (rw): USGS Optional ASCII DEM (and CDED)
  GXF (ro): GeoSoft Grid Exchange Format
  HDF5 (ro): Hierarchical Data Format Release 5
  HDF5Image (ro): HDF5 Dataset
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Christophe Lombardo
Jean-Sébastien Guay wrote:
> Is there any reference to the various formats that osgdem can use? I  
> could wait until Paul and Bob's course, but I'm curious :-)
>   
As far as I understand, osgdem relies on gdal to read dem & texture 
files. So osgdem supported formats are gdal ones and GEOTiff is part of 
them.
I am using the FWTools (http://fwtools.maptools.org/) precompiled gdal 
package instead of the one provided by osg 3rd party binaries to gain an 
access to more format and to the corresponding gdal tools (gdalinfo, 
gdalwarp an so on).

jcl




 
 

This footnote confirms that this email message has been scanned by
PineApp Mail-SeCure for the presence of malicious code, vandals & computer 
viruses.




___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Robert Osfield
On Jan 10, 2008 5:27 PM, Jean-Sébastien Guay
<[EMAIL PROTECTED]> wrote:
> Sorry for the little info, don't really know what to give. The osgdem
> without --terrain just finished, it worked and looks great. Running
> with --terrain ran a lot faster, but I can't see anything in osgviewer
> with the resulting file. The file is over twice the size (~500k versus
> ~202k without --terrain). It doesn't really matter either way - I was
> just trying it out.

Something is amiss, --terrain normally results in smaller files.  I
haven't seen any problem with rendering of them either.

Any chance you could make available the data and osgdem commndline you've used?

> Thanks, I'll look those up. Geotiffs are the most common/useful, is
> that correct? (just assuming from what I've read in the past on this
> list)

Any data with geospatial coords system assigned to them will work
pretty well, GeoTiff is the most common one for imagery.

Image and DEM's without geospatial coord system are a pain as you have
to manually specify things.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-10 Thread Jean-Sébastien Guay
Hello Robert,

> Any chance you could make available the data and osgdem commndline   
> you've used?

I've put the data at

   http://whitestar02.webhop.org/files/OSG/NED_59396523.zip

It's about 24MB. I'll remove it once you've gotten it... It comes from  
http://seamless.usgs.gov/, as instructed by the short tutorial at

   http://www.palomino3d.com/pal/openscenegraph/#GIS

(I selected a part of the western US seaboard, I think around Seattle,  
Washington)

The command lines I used were taken from that as well:

   mkdir test
   cd test
   /osgdem -d ../NED_59396523.tif -t ../NED_59396523.tif -o  
terrain.ive -a terrain.osga -l 8 -v 0.08

and

   mkdir test
   cd test
   /osgdem -d ../NED_59396523.tif -t ../NED_59396523.tif  
--terrain -o terrain.ive -a terrain.osga -l 8 -v 0.08

Let me know if that helps, or if you need something else.

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread Robert Osfield
Hi J-S,

Thanks for doing the tests.  Loading your test.osga into osgviewer on
its own it looks OK.

However, when loading your test.osga and my original test.osga into
osgviewer at the same time results the two models being loaded a long
way apart... suggesting a problem with position of the tiles.

The issue you have with all the terrain overlapping is suggest
problems with positioning of the tiles as well.  I'm currently waiting
on output.zip to download so can't say any more at this stage.

My best guess right now would be an unitialized variable or something
similiar in osgTerrain or VPB.

BTW, just how long did your full osgdem run take?  What hardware do you have?

Robert.


On Jan 11, 2008 2:58 PM, Jean-Sébastien Guay
<[EMAIL PROTECTED]> wrote:
> Hello Robert,
>
> OK, let's first try the basic terrain with one level:
>
> > The resulting test.osga is attached.  This depends on the latest
> > version of the OSG in SVN (last check in was yesterday) as it'll need
> > the latest additions to .ive for osgTerrain support.   Could you load
> > this and see what you get in the viewer.
>
> Looks fine.
>
> > Another test would be to run the above command line to see what
> > test.osga you get out, in theory is should be the same as I'm getting
> > on my machine.
> >
> >osgdem -d NED_59396523.tif -t NED_59396523.tif --terrain -o
> > test.ive -v 0.08 -a test.osga -l 1
>
> Attached, looks the same to me but not exactly the same size...
>
> Then, the more fill blown test:
>
> >  osgdem -d NED_59396523.tif -t NED_59396523.tif --terrain -o
> > output/test.ive -v 0.08
> >
> > Then to view
> >
> >   osgviewer output/test.ive
>
> Took a lot more than 40 seconds for me, but still much less than
> without --terrain...
>
> As for the result, this is a bit frustrating. What I now get is a
> bunch of tiles overlapping each other, as far as I can tell. (I used
> the exact command lines for both osgdem and osgviewer). I've attached
> a screenshot, both solid and wireframe (at extreme zoom, mind you),
> and the files (46MB zip) are at
>
>http://whitestar02.webhop.org/files/OSG/output.zip
>
> Anything else I can test? I hope this helps,
>
>
> J-S
> --
> __
> Jean-Sebastien Guay [EMAIL PROTECTED]
>  http://whitestar02.webhop.org/
>
> 
> This message was sent using IMP, the Internet Messaging Program.
>
>
> ___
> osg-users mailing list
> osg-users@lists.openscenegraph.org
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>
>
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread ümit uzun

Hi Jean; I have tried to make terrain like you from 
http://www.palomino3d.com/pal/openscenegraph/#GIS tutorial. And Fortuanetly 
create my own terrain with http://seamless.usgs.gov/ downloading some data. I 
have a unix shell which coded by Tim Brooks who are administrator of palomino3d 
web site.  I don't understand very well what is your problem in creating 
terrain but it will probably help you! Ümit UZUN > Date: Fri, 11 Jan 2008 
11:03:50 +> From: [EMAIL PROTECTED]> To: 
osg-users@lists.openscenegraph.org> Subject: Re: [osg-users] VPB in SVN> > HI 
JS,> > Thanks for the data and instructions. I have tried them and they work> 
fine on my linux box - the build runs for 40 seconds and results on a> scene 
that can be viewed with osgviewer without problem.> > One difference I use, but 
this won't effect the result, is rather than> making a sub directory, cd'ing 
into and then do a build from there is> to create the sub directory and then 
build from the original directory> i.e
 .> > mkdir output> > osgdem -d NED_59396523.tif -t NED_59396523.tif --terrain 
-o> output/test.ive -v 0.08> > Then to view> > osgviewer output/test.ive> > 
I've also tried using a .osga archive and it works fine too. FYI, The> cd'ing 
into output directly used to be required in the early days of> osgdem, but 
these days it'll automatically handle the subdirectory> specification.> > The 
fact that the database build and viewing works for for me, and not> for you, 
suggests that either VPB or OSG or both are at fault in some> way. In what way? 
I have no clue. We'll just have to step through> the various possibilities one 
by one.> > As a test I've just run the following:> > osgdem -d NED_59396523.tif 
-t NED_59396523.tif --terrain -o> test.ive -v 0.08 -a test.osga -l 1> > The 
resulting test.osga is attached. This depends on the latest> version of the OSG 
in SVN (last check in was yesterday) as it'll need> the latest additions to 
.ive for osgTerrain support. Could you load> this and se
 e what you get in the viewer.> > Another test would be to run the above 
command line to see what> test.osga you get out, in theory is should be the 
same as I'm getting> on my machine.> > Robert.> > On Jan 10, 2008 8:50 PM, 
Jean-Sébastien Guay> <[EMAIL PROTECTED]> wrote:> > Hello Robert,> >> > > Any 
chance you could make available the data and osgdem commndline> > > you've 
used?> >> > I've put the data at> >> > 
http://whitestar02.webhop.org/files/OSG/NED_59396523.zip> >> > It's about 24MB. 
I'll remove it once you've gotten it... It comes from> > 
http://seamless.usgs.gov/, as instructed by the short tutorial at> >> > 
http://www.palomino3d.com/pal/openscenegraph/#GIS> >> > (I selected a part of 
the western US seaboard, I think around Seattle,> > Washington)> >> > The 
command lines I used were taken from that as well:> >> > mkdir test> > cd test> 
> /osgdem -d ../NED_59396523.tif -t ../NED_59396523.tif -o> > terrain.ive 
-a terrain.osga -l 8 -v 0.08> >> > and> >
 > > mkdir test> > cd test> > /osgdem -d ../NED_59396523.tif -t 
 > > ../NED_59396523.tif> > --terrain -o terrain.ive -a terrain.osga -l 8 -v 
 > > 0.08> >> > Let me know if that helps, or if you need something else.> >> 
 > > >> > J-S> > --> > __> 
 > > > Jean-Sebastien Guay [EMAIL PROTECTED]> > http://whitestar02.webhop.org/> 
 > > >> > > > 
 > > This message was sent using IMP, the Internet Messaging Program.> >> >> > 
 > > ___> > osg-users mailing list> 
 > > > osg-users@lists.openscenegraph.org> > 
 > > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org> 
 > > >
_
Yeni nesil Windows Live Servisleri’ne şimdi ulaşın!
http://get.live.com#!/bin/bash
# Jim Brooks
#
# Produce an OSG scene graph from DEM and GeoTIF files
# downloaded from USGS Seamless Data Distribution System (SDDS).
# --

LEVELS=5
VERT_SCALE="0.075"
SCALE="5000.0,5000.0,5000.0"
POSITION="0,0,0"
TIF_SIZE="20% 20%"
OSGDEM_ARGS="--PagedLOD --RGB-16 --default-color 0,0,0,0 --compressed"
OSGDEM_ARGS="$OSGDEM_ARGS --no-terrain-simplification"
OSGDEM_AR

Re: [osg-users] VPB in SVN

2008-01-11 Thread Robert Osfield
On Jan 11, 2008 5:40 PM, Jean-Sébastien Guay
<[EMAIL PROTECTED]> wrote:
> Hello Robert,
>
> > The issue you have with all the terrain overlapping is suggest
> > problems with positioning of the tiles as well.  I'm currently waiting
> > on output.zip to download so can't say any more at this stage.
>
> Ok, let me know.

I get the same problem as you - once the lower levels start paging in
they all overlap each other.  This shows us at least that the problem
is in database creation rather than rendering.

Perhaps outputting the data as .osg will reveal exactly what is going amiss.

> osgdem -d NED_59396523.tif -t NED_59396523.tif --terrain -o
> output/test.ive -v 0.08
> [...]
> Elapsed time = 315.541680

Oh dear, time for you to upgrade machine and or OS.  For the complete
build  I get :

Elapsed time = 40.908958

I have a Intel 2.4GHz quad core Intel machine running Linux.   I guess
there is a chance the bugs showing themselves on the Windows side
might be affecting performance.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread Jean-Sébastien Guay
Hello Robert,

> The issue you have with all the terrain overlapping is suggest
> problems with positioning of the tiles as well.  I'm currently waiting
> on output.zip to download so can't say any more at this stage.

Ok, let me know.

> BTW, just how long did your full osgdem run take?  What hardware do you have?

I didn't time anything, mind you, and I didn't keep the output so I  
can't refer to the "Elapsed time". So I can only give an approximation  
based on my impression of how long it took - might be biased. That  
being said, I think it took about 2-3 hours.

As for the tests I did for your prevous e-mail,

osgdem -d NED_59396523.tif -t NED_59396523.tif --terrain -o test.ive  
-v 0.08 -a test.osga -l 1
[...]
Elapsed time = 0.714424

osgdem -d NED_59396523.tif -t NED_59396523.tif --terrain -o  
output/test.ive -v 0.08
[...]
Elapsed time = 315.541680

Yeah... Windows...

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread Jean-Sébastien Guay
Hello Robert,

> This is almost certainly the root of the problem with the
> visualization.  I'll add some debugging into the .ive plugin to see if
> the Locator exists in the file or whether it exists but with just
> default settings.

Any ideas as to why I didn't get the right settings? I just tried an  
update on my OSG SVN copy and nothing came in other than the  
TerrainManipulator fix Adrian sent in earlier.

> This suggest an issue with compiler as much as anything else.  It'd be
> interesting to try gcc under windows or even better the Intel
> compilers.

Possibly. I'm using the default compiler settings which are set by  
CMake. This was with a release-mode build, VS8 (VC++ 2005). I think  
this is a pretty common setup, so I'd prefer to find out what to tweak  
instead of downloading and installing MinGW (for gcc) or the ICC.

Anyways, one problem at a time... :-)

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread Jean-Sébastien Guay
Hello Robert,

> I'm stumped, I've reviewed both the DestinationTile and .ive code and
> it looks fine.  Could you put a break point in createTerrainTile() and
> step through the code that sets up the Locator to see if its setting
> up.

Sure, no problem. I'll get back to you.

> Another long shot would be to run osgdem writing to .osg rather than .ive.

I did that, it gave the same result. It has the same locator you got  
in the .osg file, i.e. default settings. So I'll try stepping in  
createTerrainTile and see.

Thanks,

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread Norman Vine
Robert Osfield writes:
> 
> On Jan 11, 2008 6:25 PM, Jean-Sébastien Guay wrote:
> > > Perhaps outputting the data as .osg will reveal exactly 
> what is going amiss.
> >
> > OK, getting closer.
> 
> I've done a
> 
> osgconv  test.ive test.osg
> 
> On your root .ive file and found the osgTerrain::Locator has default
> settings i.e.
> 
>   osgTerrain::Locator {
> CoordinateSystemType PROJECTED
> TransformScaledByResolution FALSE
> Transform {
>   1 0 0 0
>   0 1 0 0
>   0 0 1 0
>   0 0 0 1
> }
>   }
> 
> While I correctly get:
> 
> osgTerrain::Locator {
> Format "WKT"
> CoordinateSystem
> "GEOGCS[\"NAD83\",DATUM[\"North_American_Datum_1983\",SPHEROID[\"GRS
> 1980\",6378137,298.2572221010002,AUTHORITY[\"EPSG\",\"7019\"]]
> ,AUTHORITY[\"EPSG\",\"6269\"]],PRIMEM[\"Greenwich\",0],UNIT[\"
> degree\",0.0174532925199433],AUTHORITY[\"EPSG\",\"4269\"]]"
> CoordinateSystemType PROJECTED
> TransformScaledByResolution FALSE
> Transform {
>   1.14305877685547 0 0 0
>   0 0.896945953369141 0 0
>   0 0 1 0
>   -124.652221679688 41.1416664123535 0 1
> }
>   }
> 
> This is almost certainly the root of the problem with the
> visualization.  I'll add some debugging into the .ive plugin to see if
> the Locator exists in the file or whether it exists but with just
> default settings.

My guess is that the GDAL data files aren't being picked up

These are installed in a well known location on Nix  so they
Are easily picked up  this isn't necessarily so on Windows

One can tell the system where the data files are with the following 
Environment Variables

GDAL_DATA = $ROOT_OF_GDAL_DATA_FILES  # /usr/local/share/gdal  normally on
Nix
PROJ_LIB = $ROOT_OF_PROJ_DATA_FILES   # /usr/local/share/proj  normally on
Nix

HTH

Norman

___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread Robert Osfield
Hi JS,

On Jan 11, 2008 6:25 PM, Jean-Sébastien Guay
<[EMAIL PROTECTED]> wrote:
> > Perhaps outputting the data as .osg will reveal exactly what is going amiss.
>
> OK, getting closer.

I've done a

osgconv  test.ive test.osg

On your root .ive file and found the osgTerrain::Locator has default
settings i.e.

  osgTerrain::Locator {
CoordinateSystemType PROJECTED
TransformScaledByResolution FALSE
Transform {
  1 0 0 0
  0 1 0 0
  0 0 1 0
  0 0 0 1
}
  }

While I correctly get:

osgTerrain::Locator {
Format "WKT"
CoordinateSystem
"GEOGCS[\"NAD83\",DATUM[\"North_American_Datum_1983\",SPHEROID[\"GRS
1980\",6378137,298.2572221010002,AUTHORITY[\"EPSG\",\"7019\"]],AUTHORITY[\"EPSG\",\"6269\"]],PRIMEM[\"Greenwich\",0],UNIT[\"degree\",0.0174532925199433],AUTHORITY[\"EPSG\",\"4269\"]]"
CoordinateSystemType PROJECTED
TransformScaledByResolution FALSE
Transform {
  1.14305877685547 0 0 0
  0 0.896945953369141 0 0
  0 0 1 0
  -124.652221679688 41.1416664123535 0 1
}
  }

This is almost certainly the root of the problem with the
visualization.  I'll add some debugging into the .ive plugin to see if
the Locator exists in the file or whether it exists but with just
default settings.

> > I have a Intel 2.4GHz quad core Intel machine running Linux.   I guess
> > there is a chance the bugs showing themselves on the Windows side
> > might be affecting performance.
>
> That's probably the case, plus file IO and some other things might be
> slower. I have the same as you here at work (Intel quad core Q6600 @
> 2.4GHz).
>
> I have found it pretty irrelevant to compare the performance of
> offline tasks between OSes, as they're rarely comparable. Too many
> variables differ between OSes, even on the same machine. Just as an
> example, a simple raytracing task for my Masters takes about twice the
> time on Windows as on Linux on similar machines.

This suggest an issue with compiler as much as anything else.  It'd be
interesting to try gcc under windows or even better the Intel
compilers.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread Robert Osfield
Hi JS,

On Jan 11, 2008 7:36 PM, Robert Osfield <[EMAIL PROTECTED]> wrote:
> This is almost certainly the root of the problem with the
> visualization.  I'll add some debugging into the .ive plugin to see if
> the Locator exists in the file or whether it exists but with just
> default settings.

Now done this.  The test.ive that you generated contains
osgTerrain::Locator objects, and they just contain default values, so
it looks like the source of the problem is in VPB.

The code that sets up the Locator objects can be found in
DestinationTile::createTerrainTile() method in
src/vpb/Destination.cpp.  The code all looks fine, I can't spot a code
pathway that looks like it will just assign a default Locator.

I'm stumped, I've reviewed both the DestinationTile and .ive code and
it looks fine.  Could you put a break point in createTerrainTile() and
step through the code that sets up the Locator to see if its setting
up.

Another long shot would be to run osgdem writing to .osg rather than .ive.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread Jean-Sébastien Guay
Hello Norman,

> My guess is that the GDAL data files aren't being picked up

I don't think this is the problem, because the same command, without  
--terrain, produces correct output. With --terrain, it gives a Locator  
with default settings.

Still investigating. I'll continue on Monday.

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-11 Thread Jean-Sébastien Guay
Hello Robert,

> I get the same problem as you - once the lower levels start paging in
> they all overlap each other.  This shows us at least that the problem
> is in database creation rather than rendering.
>
> Perhaps outputting the data as .osg will reveal exactly what is going amiss.

OK, getting closer.

> I have a Intel 2.4GHz quad core Intel machine running Linux.   I guess
> there is a chance the bugs showing themselves on the Windows side
> might be affecting performance.

That's probably the case, plus file IO and some other things might be  
slower. I have the same as you here at work (Intel quad core Q6600 @  
2.4GHz).

I have found it pretty irrelevant to compare the performance of  
offline tasks between OSes, as they're rarely comparable. Too many  
variables differ between OSes, even on the same machine. Just as an  
example, a simple raytracing task for my Masters takes about twice the  
time on Windows as on Linux on similar machines.

J-S
-- 
__
Jean-Sebastien Guay [EMAIL PROTECTED]
 http://whitestar02.webhop.org/


This message was sent using IMP, the Internet Messaging Program.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-12 Thread Paul Martz
I haven't had time to read this entire thread from you two yet, so I
apologize if you've already discussed and resolved this...

'vsnprintf' doesn't appear to be available under VS7.1. I can't locate it in
the VS7.1 help or determine what header file should define it in VS7.1.

It builds fine with VS8, and the MSDN online help indicates it should be
defined in stdio.h, but including this in BuildLog and BuildLog.cpp for the
VS7.1 build seems to have no effect, the symbol still comes up as unknown,
and vpb fails to compile as a result.

Any ideas?

Paul Martz
Skew Matrix Software LLC
http://www.skew-matrix.com
303 859 9466

___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-12 Thread Norman Vine
Paul Martz
> 
> 'vsnprintf' doesn't appear to be available under VS7.1. I 
> can't locate it in
> the VS7.1 help or determine what header file should define it 
> in VS7.1.
> 
> Any ideas?


Hey this is Microsoft  :-)
 try  _vsnprintf

http://msdn2.microsoft.com/en-us/library/1kt27hek(VS.71).aspx

HTH

Norman

___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-15 Thread Jean-Sebastien Guay
Hello Robert,

> Still investigating. I'll continue on Monday.

OK, now with OSG and VPB updated to today's SVN, and clean builds of both, I am
getting predictable results from this command line:

osgdem -d NED_59396523.tif -t NED_59396523.tif --terrain -o output/test.ive -v
0.08

Whether it was the clean builds or the clear head (i.e. *not* Friday afternoon)
that fixed the problem, I can't really say. :-)

The only thing now is that there are lines between the tiles. See attached
screenshot. But it already looks a hundred times better than last Friday.

Thanks for your patience,

J-S
--
__
Jean-Sebastien Guay [EMAIL PROTECTED]
http://whitestar02.webhop.org/<>___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-15 Thread Robert Osfield
Hi J-S,

Good to hear things working better, although the not knowing exactly
what was amiss is a little unsettling.

The lines on the terrain are most likely down to
osgTerrain::GeometryTechnique not creating the skirts as the
old technique did.  Tomorrow I'll be adding skirts into the
GeometryTechnique and properly setting up the skirt
height from within VPB, so you'll need to update both OSG and VPB once
I've checked the changes in.

Robert.

On Jan 15, 2008 7:59 PM, Jean-Sebastien Guay
<[EMAIL PROTECTED]> wrote:
> Hello Robert,
>
> > Still investigating. I'll continue on Monday.
>
> OK, now with OSG and VPB updated to today's SVN, and clean builds of both, I 
> am
> getting predictable results from this command line:
>
> osgdem -d NED_59396523.tif -t NED_59396523.tif --terrain -o output/test.ive -v
> 0.08
>
> Whether it was the clean builds or the clear head (i.e. *not* Friday 
> afternoon)
> that fixed the problem, I can't really say. :-)
>
> The only thing now is that there are lines between the tiles. See attached
> screenshot. But it already looks a hundred times better than last Friday.
>
> Thanks for your patience,
>
>
> J-S
> --
> __
> Jean-Sebastien Guay [EMAIL PROTECTED]
> http://whitestar02.webhop.org/
> ___
> osg-users mailing list
> osg-users@lists.openscenegraph.org
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
>
>
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-15 Thread Jean-Sebastien Guay
Hello Robert,

> Good to hear things working better, although the not knowing exactly
> what was amiss is a little unsettling.

Agreed. I assure you I'm on the lookout for any further signs that things are
wonky on my side... :-)

> The lines on the terrain are most likely down to
> osgTerrain::GeometryTechnique not creating the skirts as the
> old technique did.  Tomorrow I'll be adding skirts into the
> GeometryTechnique and properly setting up the skirt
> height from within VPB, so you'll need to update both OSG and VPB once
> I've checked the changes in.

OK. Do you get the same behaviour on your end?

J-S
--
__
Jean-Sebastien Guay [EMAIL PROTECTED]
http://whitestar02.webhop.org/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-15 Thread Robert Osfield
On Jan 15, 2008 8:55 PM, Jean-Sebastien Guay
<[EMAIL PROTECTED]> wrote:
> OK. Do you get the same behaviour on your end?

Yes, the gaps between tiles of different levels is usual right now.
The --terrain feature is very new to VPB, its not yet productive ready
so one would expect little items to not work 100%.  It'll be working
fully pretty soon though ;-)

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-15 Thread Jean-Sebastien Guay
Hello Robert,

> Yes, the gaps between tiles of different levels is usual right now.
> The --terrain feature is very new to VPB, its not yet productive ready
> so one would expect little items to not work 100%.  It'll be working
> fully pretty soon though ;-)

It's not just different levels though - all tiles seem to have black borders.
Even when I'm zoomed all the way in and all tiles are at the highest LOD, there
are still those black lines.

And I totally understand about --terrain. Without knowing the details, I'd say
that faster generation is always welcome! I'm just giving feedback to make sure
it's not something in my setup or something platform-specific.

J-S
--
__
Jean-Sebastien Guay [EMAIL PROTECTED]
http://whitestar02.webhop.org/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-16 Thread Robert Osfield
Hi J-S,

On Jan 16, 2008 12:11 AM, Jean-Sebastien Guay
<[EMAIL PROTECTED]> wrote:
> > Yes, the gaps between tiles of different levels is usual right now.
> > The --terrain feature is very new to VPB, its not yet productive ready
> > so one would expect little items to not work 100%.  It'll be working
> > fully pretty soon though ;-)
>
> It's not just different levels though - all tiles seem to have black borders.
> Even when I'm zoomed all the way in and all tiles are at the highest LOD, 
> there
> are still those black lines.

O.K. this isn't the lack of skirts then.

> And I totally understand about --terrain. Without knowing the details, I'd say
> that faster generation is always welcome! I'm just giving feedback to make 
> sure
> it's not something in my setup or something platform-specific.

Feedback is useful, especially when it uncovers problems that I not
able to reproduce myself which does seem to be the case here.

A black line around the border suggests either a source texture,
texture coordinate issue or normal generation issue.  The later two
possibilities would be related to the core OSG, the first would be an
osgdem issue.  Could you package up you model again and make it
available for download so I can at least discount/confirm the source
texture as being at fault.

Robert.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-16 Thread Jean-Christophe Lombardo
Hi,

I have exactly the same black tile border problem on windows side with 
my own data set.
On linux side it seems ok.

jcl




 
 

This footnote confirms that this email message has been scanned by
PineApp Mail-SeCure for the presence of malicious code, vandals & computer 
viruses.




___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-16 Thread Jean-Christophe Lombardo
Just to go a little further on this black contour problem:
I've transfered my scene generated on linux side to windows, and the 
black contour appears...

Is there any difference in the default texture parameters between linux 
and windows ?

Thanks

jcl



 
 

This footnote confirms that this email message has been scanned by
PineApp Mail-SeCure for the presence of malicious code, vandals & computer 
viruses.




___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-16 Thread Robert Osfield
On Jan 16, 2008 1:12 PM, Jean-Christophe Lombardo
<[EMAIL PROTECTED]> wrote:
> Just to go a little further on this black contour problem:
> I've transfered my scene generated on linux side to windows, and the
> black contour appears...
>
> Is there any difference in the default texture parameters between linux
> and windows ?

Shouldn't be an difference.  Perhaps its the OpenGL drivers show small
variations of interpretation of the same settings.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-16 Thread Jean-Sebastien Guay
Hello Robert,

> I've been comparing the texture setup up code in both VPB and
> osgTerrain and found that VPB uses CLAMP_TO_EDGE while
> osgTerrain::GeometryTechnique itself just used defaults.  I've changed
> osgTerrain::GeometryTechnqiue to use CLAMP_TO_EDGE, testing under
> linux doesn't show any difference, but there is a chance that it might
> change things under Windows.  The change is now checked in.

Just updated OSG from SVN to get the changes to osgTerrain::GeometryTechnique,
recompiled, and regenerated the terrain, and it's fixed. No more black lines.
Thanks Robert.

I don't think there are any other issues right now, unless you want me to
download some other data set and do some tests on it... The next step seems to
be documenting and testing the distributed database generation.

Thanks for your help,

J-S
--
__
Jean-Sebastien Guay [EMAIL PROTECTED]
http://whitestar02.webhop.org/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-16 Thread Robert Osfield
On Jan 16, 2008 2:51 PM, Jean-Sebastien Guay
<[EMAIL PROTECTED]> wrote:
> Just updated OSG from SVN to get the changes to osgTerrain::GeometryTechnique,
> recompiled, and regenerated the terrain, and it's fixed. No more black lines.
> Thanks Robert.

Good to hear.

> I don't think there are any other issues right now, unless you want me to
> download some other data set and do some tests on it... The next step seems to
> be documenting and testing the distributed database generation.

Next step is for me to complete the distributed functionality, then
I'll have a bit of breathing space to document it.

Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-16 Thread Paul Martz
NVIDIA drivers have a control for default texture clamping. I know this used
to default to the CLAMP_TO_EDGE, and if you wanted OpenGL spec-compliant
behavior, you had to manually change it to CLAMP_TO_BORDER.

I wonder if this is the cause of the black lines issue? CLAMP_TO_BORDER with
full 0 to 1 texture coordinates and no border specified in the texture, that
would cause a black edge where the texture tiles meet... If this is the
case, proper use of the texture border would eliminate the problem.

Paul Martz
Skew Matrix Software LLC
http://www.skew-matrix.com
303 859 9466



> -Original Message-
> From: [EMAIL PROTECTED] 
> [mailto:[EMAIL PROTECTED] On Behalf 
> Of Jean-Christophe Lombardo
> Sent: Wednesday, January 16, 2008 6:13 AM
> To: OpenSceneGraph Users
> Subject: Re: [osg-users] VPB in SVN
> 
> Just to go a little further on this black contour problem:
> I've transfered my scene generated on linux side to windows, 
> and the black contour appears...
> 
> Is there any difference in the default texture parameters 
> between linux and windows ?
> 
> Thanks
> 
> jcl
> 
> 
> 
>  
>  
> **
> **
> This footnote confirms that this email message has been 
> scanned by PineApp Mail-SeCure for the presence of malicious 
> code, vandals & computer viruses.
> **
> **
> 
> 
> 
> ___
> osg-users mailing list
> osg-users@lists.openscenegraph.org
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-opensce
negraph.org

___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] VPB in SVN

2008-01-16 Thread Jean-Sebastien Guay
Hi Paul,

> NVIDIA drivers have a control for default texture clamping. I know this used
> to default to the CLAMP_TO_EDGE, and if you wanted OpenGL spec-compliant
> behavior, you had to manually change it to CLAMP_TO_BORDER.

Thanks for the details! I'm always impressed at how difficult it seems to be for
driver writers to conform to a spec... I'm sure it's more difficult than it
seems, but I wonder if they're really trying hard enough?

> I wonder if this is the cause of the black lines issue? CLAMP_TO_BORDER with
> full 0 to 1 texture coordinates and no border specified in the texture, that
> would cause a black edge where the texture tiles meet... If this is the
> case, proper use of the texture border would eliminate the problem.

Yep, that's whay Robert ended up doing.

Thanks,

J-S
--
__
Jean-Sebastien Guay [EMAIL PROTECTED]
http://whitestar02.webhop.org/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org