[PD] GEM on PIII memory problem?

IOhannes m zmoelnig zmoelnig at iem.at
Tue Jan 17 09:15:51 CET 2012

Hash: SHA1

On 2012-01-16 18:06, rolf meesters wrote:
> Subject: Re: [PD] GEM on PIII memory problem?
>> is that with the intel gfx card?
> the PIII has onboard video

which is?

>> could it be a dupe of [1]?
> not as far as i can see. (i use fresh copies)

you are using a recent git checkout of Gem?

> i suspect [1] has something to do with using the pd open or close 'message'
> with a non-existing file.

oh indeed. that was the wrong bug report i mentioned. i wanted to refer
to #3471122 ("GL: invalid enum" in each frame); but if you are using a
recent version this should not be the problem....

> or what does "every time at a change in the rendering" mean exactly?
> every time the content of the GEM-window should change

i assume that you mean, whenever you rotate/translate/scale/... an
object in the scene?
there are so many ways you can make he GEM-window change. to name a few:
- - whenever you "destroy, create" a gemwin
- - whenever you change global ligthing
- - whenever you move things in the 3d scene
- - whenever someone moves in front of the camera that is captured and
applied as a texture
- - ....

those changes are very different and might hint at different problems.
no doubt it is very clear to you what you mean by "the content should
change"; however, it is not so clear to me, that's why i keep asking.

if you can provide a minimal patch that triggers the problem, this might
greatly improve my understanding.

>> try to find the object that actually causes the error.
>> you can use a [GEMglReportError] that will output a warning if an error
>> was raised since the last time any instance of [GEMglReportError] was
>> called (the error state is also reset inside [gemwin] during each
>> render-cycle..which is what causes the message to appear)
> it's not clear to me how to use this object, there's no help.
> connecting the inlet with a loadbang causes Pd/GEM to quit immediately.
> leaving the object by itself, with a print on the first outlet, gives
> nothing.

right, sorry for that.
[GEMglReportError] should be part of the render-chain.

here's an example:


where you suspect that [weirdobject] is the cause of the problem.
if this is true, every time [weirdobject] does something stupid (in
openGL world), the lower [GEMglReportError] will send a non-null error
code to it's right outlet.
the other (upper) [GEMglReportError] is needed to clear any previous
error (in case you have multiple objects doing stupid things)

if things are very weird, you might even use something like

[t     a      a       a]
|                   |  |
[GEMglReportError]  |  [GEMglReportError]

> should i make a bug report on this? (i'm on Pd-ext 42.5 with GEM 92.3)

ahem, so which version of gem are you actually using?
the current stable release is 0.93.3, and i was under the impression
that you were using a "fresh copy" of Gem.


Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/


-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3636 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20120117/bac20f93/attachment.bin>

More information about the Pd-list mailing list