Here at work, I am the only one interested into the new itools, yet
when running idl (linux) I get the following errors and the axis are
unreadable.
Did anyone have the same problem?

IDL> iplot,findgen(100)
% Loaded DLM: PNG.
% Compiled module: XMANAGER.
% Loaded DLM: IPTOOL.
% Compiled module: GET_SCREEN_SIZE.
% Compiled module: REVERSE.
% Compiled module: CVTTOBM.
% Compiled module: IDENTITY.
% Compiled module: STRSPLIT.
% Compiled module: UNIQ.
IDL> Failed to upload texture, sz 128
                                     Memory heap (nil):
                                                         heap == 0
                                                                  End
of memory blocks
      Failed to upload texture, sz 512
                                      Memory heap (nil):
                                                          heap == 0
                                                                   End
of memory blocks


----

Thijs Krijger

krijger@astro.uu.nl writes:

> Here at work, I am the only one interested into the new itools, yet
> when running idl (linux) I get the following errors and the axis are
> unreadable.
> Did anyone have the same problem?

The *only* one!? Gosh, it's worse than I thought. :-(

I'd contact RSI. This looks like an installation
problem to me. You might try setting your graphics
renderer to "software".

Cheers,

David

-- 
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155

<krijger@astro.uu.nl> wrote in message
news:3fb11ca9.0310090542.1f8ab30@posting.google.com...
> Here at work, I am the only one interested into the new itools, yet
> when running idl (linux) I get the following errors and the axis are
> unreadable.
> Did anyone have the same problem?
>
> IDL> iplot,findgen(100)
> % Loaded DLM: PNG.
> % Compiled module: XMANAGER.
> % Loaded DLM: IPTOOL.
> % Compiled module: GET_SCREEN_SIZE.
> % Compiled module: REVERSE.
> % Compiled module: CVTTOBM.
> % Compiled module: IDENTITY.
> % Compiled module: STRSPLIT.
> % Compiled module: UNIQ.
> IDL> Failed to upload texture, sz 128
>                                      Memory heap (nil):
>                                                          heap == 0
>                                                                   End
> of memory blocks
>       Failed to upload texture, sz 512
>                                       Memory heap (nil):
>                                                           heap == 0
>                                                                    End
> of memory blocks
>

I checked, and I don't think that IDL or iTools is capable of generating
these messages.  My best guess is that these are some sort of debugging or
error messages coming out of your linux OpenGL driver.  David's advice is
good; whenever you suspect a graphics driver problem, try using software
rendering.  I would also look really hard for an upgrade for your drivers.
If the drivers are bad enough that you'd like to abandon hardware rendering
completely for IDL on your system, find the gl_driver.so module in your IDL
binary directory and move/rename it.

A couple of things changed in IDL 6.0 that may be of interest here.  The
Object Graphics IDLgrText object now renders text by using texture mapping,
and that's why the innocent-looking iPlot command really does use textures,
even though there isn't a texture-mapped object like a surface in sight.
This will tend to exercise the texture mapping hardware and software in an
OpenGL system more than it used to, and that's part of what might be
exposing this problem in your driver.  You might want to try some simpler
tests like "XOBJVIEW, OBJ_NEW('IDLgrAxis')"  or "XOBJVIEW,
OBJ_NEW('IDLgrText', "Hello", RENDER_METHOD=0)" and play with different
values of RENDER_METHOD (0=texture, 1=triangles), just to try to learn more
about the problem.

IDL 6.0 on linux also is now more likely to take advantage of
hardware-accelerated OpenGL.  Although the entire hardware GL on linux story
has gotten much better in the past year, there are still some
"bleeding-edge" drivers out there that are not quite mature yet.  We've had
good results here at RSI with a few configurations, such as nVidia and ATI
drivers on RedHat 8 and 9 platforms.  But in a lot of cases with newer
drivers, you're not always guaranteed to get results as good as we have
seen.

It might be helpful if you could post more details about your configuration,
including graphics hardware, linux distro and version, XFree86 version, and
OpenGL driver version.  Perhaps someone else has already worked through
this.

Karl

David Fanning wrote:
> krijger@astro.uu.nl writes:
> 
> 
>>Here at work, I am the only one interested into the new itools, yet
>>when running idl (linux) I get the following errors and the axis are
>>unreadable.
>>Did anyone have the same problem?
> 
> 
> The *only* one!? Gosh, it's worse than I thought. :-(
> 

That's one more than here.

The only new feature I was really interested in was the IDL Java bridge. 
Unfortunately I just can't get it to work properly on Linux, and they've not 
even bothered to implement it on IRIX.

So, for now, IDL 6.0 lies unused...


-- 
Nigel Wade, System Administrator, Space Plasma Physics Group,
             University of Leicester, Leicester, LE1 7RH, UK
E-mail :    nmw@ion.le.ac.uk
Phone :     +44 (0)116 2523548, Fax : +44 (0)116 2523555

Nigel Wade writes:

> > The *only* one!? Gosh, it's worse than I thought. :-(
> > 
> 
> That's one more than here.

The only one in Europe, is what I meant. :-)

Cheers,

David

-- 
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155

> You might want to try some simpler tests like 
> "XOBJVIEW, OBJ_NEW('IDLgrAxis')" 

Same error.

> "XOBJVIEW,OBJ_NEW('IDLgrText', "Hello", RENDER_METHOD=0)" 

Same error

> and play with different
> values of RENDER_METHOD (0=texture, 1=triangles), just to try to learn more
> about the problem.

> "XOBJVIEW,OBJ_NEW('IDLgrText', "Hello", RENDER_METHOD=1)" 

works

> It might be helpful if you could post more details about your configuration,
> including graphics hardware, linux distro and version, XFree86 version, and
> OpenGL driver version.  Perhaps someone else has already worked through
> this.

Matrox 450
Redhad 7.3
Xfree 4.2.1
OPenGL 1.3

---
Thijs

<krijger@astro.uu.nl> wrote in message
news:3fb11ca9.0310130107.307d36c2@posting.google.com...
> > You might want to try some simpler tests like
> > "XOBJVIEW, OBJ_NEW('IDLgrAxis')"
>
> Same error.
>
> > "XOBJVIEW,OBJ_NEW('IDLgrText', "Hello", RENDER_METHOD=0)"
>
> Same error
>
> > and play with different
> > values of RENDER_METHOD (0=texture, 1=triangles), just to try to learn
more
> > about the problem.
>
> > "XOBJVIEW,OBJ_NEW('IDLgrText', "Hello", RENDER_METHOD=1)"
>
> works

Between this and the original error messages, I'm pretty sure that your
driver is broken with respect to texture mapping.

I would switch IDL to software rendering as I noted in my previous post and
start looking for driver fixes.

The only other thing that might be worth doing is to find a really simple
stand-alone OpenGL sample program that does texture mapping and see if that
runs.  If it does not, then you have a simple test case that reproduces the
problem which is good to have if you report the problem to Matrox or whoever
is working on the drivers.

>
> > It might be helpful if you could post more details about your
configuration,
> > including graphics hardware, linux distro and version, XFree86 version,
and
> > OpenGL driver version.  Perhaps someone else has already worked through
> > this.
>
> Matrox 450
> Redhad 7.3
> Xfree 4.2.1
> OPenGL 1.3
>
> ---
> Thijs