wow,
i just had a full crash on my otherwise ok-running gentoo station when running pd0.37+pdp+pmpd+pidip. (the oscar goes to the journalling file system).
there was nothing in the logs, yet what i used was
- parts of the example 02 of pmpd
- standard pdp stuff that usually does not crash on this machine
- [pdp_capture] from pidip - never used before
the incident happened right when i was dragging some objects in the patcher.
i would really like to know, what object is to blame ... all of the libs are fairly new.
anyone ever experienced a hard crash with one of the above? i mean one with locked keyboard and null happening afterwards.
lg martin
attacksyour.net/pi
anyone ever experienced a hard crash with one of the above? i mean one with locked keyboard and null happening afterwards.
lg martin
yes, you are not alone. But I think it is a GEM problem, and has nothing to do with Yves objects. I managed to crash my debian box with the internal Gem objects alone. I could not reproduce these errors, I think it could have been a misconfiguration (wrong graphic driver) of the system or libraries compiled with different versions of gcc???
marius.
hi
i am on os X 10.2.8 attempting to load the freeverb~.pd darwin external but whilst loading it fails with
dyld: ./pd Undefined symbols: ___gxx_personality_v0 pd_gui: pd process exited Trace/BPT trap
can someone help with this?
cheers
Patrick
On Mon, 10 May 2004, shreeswifty wrote:
i am on os X 10.2.8 attempting to load the freeverb~.pd darwin external but whilst loading it fails with dyld: ./pd Undefined symbols: ___gxx_personality_v0 can someone help with this?
Afaik, gcc's personality problems are due to it not really be like g++. But for C++ you have to use "g++" instead of "gcc". Why is that, I don't know, you can compile C++ code with "gcc", which is supposed to be a unified frontend, it's just that it doesn't run. :-(
Mathieu Bouchard http://artengine.ca/matju
marius schebella wrote:
anyone ever experienced a hard crash with one of the above? i mean one with locked keyboard and null happening afterwards.
lg martin
yes, you are not alone. But I think it is a GEM problem, and has nothing to do with Yves objects.
gracias, i couldn't say my objects never ever crash, especially when it gets to video processing, but i experienced some freezes too with Gem with that new kernel 2.6.
it did not use to happen with kernel 2.5 or 2.4.
and the pdp_capture is terribly slow and eats a lot of ressources, that is true, as well as pdp2gem because it's using capture techniques that are quite heavy for the graphic driver.
cheers, sevy
hi.
here comes the lonesome defender of Gem...
Yves Degoyon wrote:
yes, you are not alone. But I think it is a GEM problem, and has nothing to do with Yves objects.
i wouldn't have guessed this from martin's email that was explicitly talking about pd0.37+pdp+pmpd+pidip (i guess (at least: i hope) he would have mentioned Gem if it were in the list)
gracias, i couldn't say my objects never ever crash, especially when it gets to video processing,
but i experienced some freezes too with Gem with that new kernel 2.6.
it did not use to happen with kernel 2.5 or 2.4.
this is true. as far asi have tracked it down, it is really (honestly ;-)) related to QuickTime (and probably the combination QuickTime+openGL ??) i haven't tried pdp/pidip, but i guess it works ok here. anyhow, my reference is the "lqtplay"-application that is coming with the quicktime-utilities (so I haven't written a single line of this code) and it religiously (as chris would say) freezes with kernel-2.6(.3) - exactly like Gem.
anyhow, these crashes are normally rather lightweight, as you could just kill pd and restart. (no keyboard freeze)
mfg.asd.r IOhannes
hi,
i wouldn't have guessed this from martin's email that was explicitly talking about pd0.37+pdp+pmpd+pidip (i guess (at least: i hope) he would have mentioned Gem if it were in the list)
you are right - gem was not loaded.
but i experienced some freezes too with Gem with that new kernel 2.6. it did not use to happen with kernel 2.5 or 2.4.
this is true. as far asi have tracked it down, it is really (honestly ;-)) related to QuickTime (and probably the combination QuickTime+openGL ??) i haven't tried pdp/pidip, but i guess it works ok here. anyhow, my reference is the "lqtplay"-application that is coming with the quicktime-utilities (so I haven't written a single line of this code) and it religiously (as chris would say) freezes with kernel-2.6(.3) - exactly like Gem.
actually i am using 2.4.22 exactly because of this (and because my measurements show that i runs faster than the new kernel when it comes to graphic processing on this computer).
for sure, the reasons for that crash cannot be found by discussing it on the list. i was just so upset since i had not saved for some minutes and what i tried to do looked good for the first time. you know the process. maybe it was just a ruined quicktime movie....
but it could have been that someone experienced the same.
lg martin
attacksyour.net/pi
martin pichlmair wrote:
hi,
i wouldn't have guessed this from martin's email that was explicitly talking about pd0.37+pdp+pmpd+pidip (i guess (at least: i hope) he would have mentioned Gem if it were in the list)
you are right - gem was not loaded.
ooops, didn't read that before answering Iohannes mail.
sorry, ok, it's not a Gem problem nor a 2.6 kernel problem, then, uhhh, i dunno, try to provide more informations.
cheers, sevy
IOhannes m zmoelnig wrote:
hi.
here comes the lonesome defender of Gem...
Yves Degoyon wrote:
yes, you are not alone. But I think it is a GEM problem, and has nothing to do with Yves objects.
i wouldn't have guessed this from martin's email that was explicitly talking about pd0.37+pdp+pmpd+pidip (i guess (at least: i hope) he would have mentioned Gem if it were in the list)
ok, iohannes, not blaming Gem too much, i know libquicktime has a lot of problem and maybe more with recent kernels.
still, Gem was in the list of what was used there, since pmpd is Gem's based.
but, one thing, i don't think it uses libquicktime, it's only synthesis from what i know.
the best thing is maybe is maybe to isolate the problem by using less components ( and maybe this is due to pdp too, we're just working on suppositions here ).
cheers, sevy
-
IOhannes m zmoelnig -
marius schebella -
martin pichlmair -
Mathieu Bouchard -
rat＠telecoma.net -
shreeswifty -
Yves Degoyon