From: Andreas K. H. <and...@ph...> - 2009-10-17 20:26:31
|
Hi, here's some more info: I downloaded the wrml file from http://www.ch.ic.ac.uk/rzepa/vrml/brickmann/panel1b.wrl > Before you start up openvrml-player or the browser plug-in, start up > openvrml-xembed in another terminal window. The front end should use > that process instead of starting a new one. > Do you get any console output? Yes: from the player: huettel@pinacolada ~ $ openvrml-player ** (openvrml-player:28730): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed ** (openvrml-player:28730): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed ** (openvrml-player:28730): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed (and this repeated many times) and from the xembed process huettel@pinacolada ~ $ /usr/libexec/openvrml-xembed ** (openvrml-xembed:28728): DEBUG: erased references to :1.1885 ** (openvrml-xembed:28728): DEBUG: erased references to :1.1889 ** (openvrml-xembed:28728): DEBUG: erased references to :1.1889 ** (openvrml-xembed:28728): DEBUG: erased references to :1.1890 ** (openvrml-xembed:28728): DEBUG: erased references to :1.1891 ** (openvrml-xembed:28728): DEBUG: erased references to :1.1879 Thanks a lot for any advice, best, Andreas -- Dr. Andreas K. Huettel Institute for Experimental and Applied Physics University of Regensburg D-93040 Regensburg Germany tel. +49 151 241 67748 (mobile) e-mail ma...@ak... http://www.akhuettel.de/research/ |
From: Braden M. <br...@en...> - 2009-10-18 06:10:17
|
On Sat, 2009-10-17 at 22:28 +0200, Andreas K. Huettel wrote: > Hi, > > here's some more info: I downloaded the wrml file from > http://www.ch.ic.ac.uk/rzepa/vrml/brickmann/panel1b.wrl > > > Before you start up openvrml-player or the browser plug-in, start up > > openvrml-xembed in another terminal window. The front end should use > > that process instead of starting a new one. > > > Do you get any console output? > > Yes: from the player: > > huettel@pinacolada ~ $ openvrml-player > > ** (openvrml-player:28730): CRITICAL **: > void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi > handle' failed > > ** (openvrml-player:28730): CRITICAL **: > void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi > handle' failed > > ** (openvrml-player:28730): CRITICAL **: > void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi > handle' failed > > (and this repeated many times) I can reproduce this. Before I can run into bugs in OpenVRML, there are a couple of predicate issues: * If you downloaded this file and tried to load it from your local machine, the relative URIs to the Inline'd worlds would be wrong. You'd need to download those, too. * The Inline'd worlds appear to be gzipped. This would be handled by libcurl (for openvrml-player) or the Web browser (for the browser plug-in) if the server were sending the correct Content-Encoding; but it doesn't appear to be doing that. Now, once I got one of the Inline'd worlds (and commented out the rest) and gunzipped it, I still ran into problems. I haven't dug into this much yet, but it looks like there must be some significant problem with Inline. I'll investigate further. > and from the xembed process > > huettel@pinacolada ~ $ /usr/libexec/openvrml-xembed > ** (openvrml-xembed:28728): DEBUG: erased references to :1.1885 > ** (openvrml-xembed:28728): DEBUG: erased references to :1.1889 > ** (openvrml-xembed:28728): DEBUG: erased references to :1.1889 > ** (openvrml-xembed:28728): DEBUG: erased references to :1.1890 > ** (openvrml-xembed:28728): DEBUG: erased references to :1.1891 > ** (openvrml-xembed:28728): DEBUG: erased references to :1.1879 You should see an "inserted reference" when you start openvrml-player (or the browser plug-in) when things are working correctly. I'm going to assume, for the moment, that this was overlooked; because I don't think you could have gotten as far as getting those error messages from openvrml-player if this hadn't happened. You might want to try a very simple VRML file to sanity check the basics. -- Braden McDaniel <br...@en...> |
From: Andreas K. H. <and...@ph...> - 2009-10-18 17:07:47
|
Am Sonntag 18 Oktober 2009 08:10:02 schrieb Braden McDaniel: > I can reproduce this. Before I can run into bugs in OpenVRML, there are > a couple of predicate issues: > > * If you downloaded this file and tried to load it from your local > machine, the relative URIs to the Inline'd worlds would be > wrong. You'd need to download those, too. Point taken... > You should see an "inserted reference" when you start openvrml-player > (or the browser plug-in) when things are working correctly. Up to now there never was such a message. I tried something else, I pasted the http URL of the file into the address field of the player. This is the result: Output of xembed (in the end I killed it with ctrl-c): huettel@pinacolada ~ $ /usr/libexec/openvrml-xembed ** (openvrml-xembed:6417): DEBUG: inserted reference to :1.2098 ** ERROR **: Out of memory aborting... ** ERROR **: Out of memory aborting... Abgebrochen huettel@pinacolada ~ $ Output of the player: a few hundred :) ** (openvrml-player:6424): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed > > You might want to try a very simple VRML file to sanity check the > basics. > Next test with downloaded copy of http://tecfa.unige.ch/guides/vrml/examples/basics/basics-1.wrl -> seems to work, displays the shapes and I can rotate them Next test with this URL directly in the address bar: -> seems to work fine too In both cases no output of the player and the xembed shows huettel@pinacolada ~ $ /usr/libexec/openvrml-xembed ** (openvrml-xembed:6535): DEBUG: inserted reference to :1.2113 ** (openvrml-xembed:6535): DEBUG: erased references to :1.2113 huettel@pinacolada ~ $ The browser plugin also displays the example correctly. So this seems to be a problem of "downloading embedded references"? Would make sense since the error messages are from curl... Best, Andreas -- Dr. Andreas K. Huettel Institute for Experimental and Applied Physics University of Regensburg D-93040 Regensburg Germany tel. +49 151 241 67748 (mobile) e-mail ma...@ak... http://www.akhuettel.de/research/ |
From: Braden M. <br...@en...> - 2009-10-19 16:44:43
|
On Sun, 2009-10-18 at 19:09 +0200, Andreas K. Huettel wrote: > Am Sonntag 18 Oktober 2009 08:10:02 schrieb Braden McDaniel: > > > I can reproduce this. Before I can run into bugs in OpenVRML, there are > > a couple of predicate issues: > > > > * If you downloaded this file and tried to load it from your local > > machine, the relative URIs to the Inline'd worlds would be > > wrong. You'd need to download those, too. > > Point taken... > > > You should see an "inserted reference" when you start openvrml-player > > (or the browser plug-in) when things are working correctly. > > Up to now there never was such a message. Well, if openvrml-xembed never prints an "inserted reference" message to the console, you have other problems: i.e., that means that no control has been created for the host--which probably means that the host never found the openvrml-xembed process. > I tried something else, I pasted the http URL of the file into the address > field of the player. This is the result: > > Output of xembed (in the end I killed it with ctrl-c): > > huettel@pinacolada ~ $ /usr/libexec/openvrml-xembed > ** (openvrml-xembed:6417): DEBUG: inserted reference to :1.2098 > > ** ERROR **: Out of memory > aborting... > > ** ERROR **: Out of memory > aborting... > Abgebrochen > huettel@pinacolada ~ $ > > Output of the player: a few hundred :) > > ** (openvrml-player:6424): CRITICAL **: > void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi > handle' failed This may be a bit simpler than I thought. I think openvrml-player starts producing these messages once openvrml-xembed becomes unresponsive (i.e., it crashed). That would happen because a bug in the Collision node implementation is causing infinite recursion and blowing out the stack. I've now fixed that on the trunk and branches/0.18. Can you give it a try? -- Braden McDaniel <br...@en...> |
From: Andreas K. H. <and...@ph...> - 2009-10-21 11:27:07
|
I am using trunk now (r 4006). http://www.ch.ic.ac.uk/rzepa/vrml/brickmann/panel1b.wrl this problem V did not change at all > I tried something else, I pasted the http URL of the file into the address > field of the player. This is the result: > > Output of xembed (in the end I killed it with ctrl-c): > > huettel@pinacolada ~ $ /usr/libexec/openvrml-xembed > ** (openvrml-xembed:6417): DEBUG: inserted reference to :1.2098 > > ** ERROR **: Out of memory > aborting... > > ** ERROR **: Out of memory > aborting... > Abgebrochen > huettel@pinacolada ~ $ > > Output of the player: a few hundred :) > > ** (openvrml-player:6424): CRITICAL **: > void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi > handle' failed > I downloaded all the subfiles of this wrl, gunzipped them and placed them in the same directory. I stll get the errors * for the xembed: huettel@grenadine ~ $ /usr/libexec/openvrml-xembed ** (openvrml-xembed:18283): DEBUG: inserted reference to :1.305 ** ERROR **: Out of memory aborting... Abgebrochen huettel@grenadine ~ $ * for the player: huettel@grenadine ~ $ openvrml-player & [1] 18285 huettel@grenadine ~ $ ** (openvrml-player:18285): CRITICAL **: Call to org.openvrml.Browser.NewStream failed: Message did not receive a reply (timeout by message bus) ** (openvrml-player:18285): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed ** (openvrml-player:18285): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed ** (openvrml-player:18285): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed ... Very simple example files seem to work... Best, Andreas -- Dr. Andreas K. Huettel Institute for Experimental and Applied Physics University of Regensburg D-93040 Regensburg Germany tel. +49 151 241 67748 (mobile) e-mail ma...@ak... http://www.akhuettel.de/research/ |
From: Andreas K. H. <and...@ph...> - 2009-10-21 12:39:16
|
PS. If I make a copy of the main wrl file and replace all Inline { url "..." } blocks with the contents of the respective file, the model is displayed correctly! > I downloaded all the subfiles of this wrl, gunzipped them and placed them in the same directory. I stll get the errors > > * for the xembed: > > huettel@grenadine ~ $ /usr/libexec/openvrml-xembed > ** (openvrml-xembed:18283): DEBUG: inserted reference to :1.305 > > ** ERROR **: Out of memory > aborting... > Abgebrochen > huettel@grenadine ~ $ > > > * for the player: > > huettel@grenadine ~ $ openvrml-player & > [1] 18285 > huettel@grenadine ~ $ > ** (openvrml-player:18285): CRITICAL **: Call to org.openvrml.Browser.NewStream failed: Message did not receive a reply (timeout by message bus) > > ** (openvrml-player:18285): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed > > ** (openvrml-player:18285): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed > > ** (openvrml-player:18285): CRITICAL **: void<unnamed>::reset_fds(<unnamed>::CURLSource&): assertion `Invalid multi handle' failed > > ... > > > Very simple example files seem to work... > > > Best, Andreas > > > > > > > > > > > -- Dr. Andreas K. Huettel Institute for Experimental and Applied Physics University of Regensburg D-93040 Regensburg Germany tel. +49 151 241 67748 (mobile) e-mail ma...@ak... http://www.akhuettel.de/research/ |
From: Braden M. <br...@en...> - 2009-10-22 06:00:44
|
On Wed, 2009-10-21 at 13:04 +0200, Andreas K. Huettel wrote: > I am using trunk now (r 4006). > > http://www.ch.ic.ac.uk/rzepa/vrml/brickmann/panel1b.wrl > this problem V did not change at all Well, there is still a problem... but the nature of it has definitely changed. Sometimes I get output like you posted; and sometimes my output from trunk openvrml-xembed looks like this: $ OPENVRML_DATADIR=$(pwd)/data OPENVRML_NODE_PATH=$(pwd)/build/src/node OPENVRML_SCRIPT_PATH=$(pwd)/build/src/script ./build/src/openvrml-xembed/openvrml-xembed ** (lt-openvrml-xembed:680): DEBUG: inserted reference to :1.560 file:///home/braden/Models/dna_w.wrl: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken. no alternative URI could be resolved unreachable URI file:///home/braden/Models/hot_cst.wrl: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken. no alternative URI could be resolved unreachable URI file:///home/braden/Models/Sc2_esp.wrl: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken. no alternative URI could be resolved unreachable URI lt-openvrml-xembed: .././../src/local/libopenvrml-control/openvrml_control/browser.cpp:705: void openvrml_control::browser::new_stream(uint64_t, const std::string&, const std::string&): Assertion `this->uninitialized_streambuf_map_.size() == 1' failed. And if I comment out all but two of the Inlines, sometimes I can even get it to work. Significantly, the file seems to work fine in sdl-viewer. So there's something (or things) amiss with stream handling between the host and the control. I've entered this in trac here: https://sourceforge.net/apps/trac/openvrml/ticket/92 -- Braden McDaniel <br...@en...> |
From: Braden M. <br...@en...> - 2010-02-06 05:40:32
|
On Thu, 2009-10-22 at 02:00 -0400, Braden McDaniel wrote: > On Wed, 2009-10-21 at 13:04 +0200, Andreas K. Huettel wrote: > > I am using trunk now (r 4006). > > > > http://www.ch.ic.ac.uk/rzepa/vrml/brickmann/panel1b.wrl > > this problem V did not change at all [snip] > So there's something (or things) amiss with stream handling between the > host and the control. > > I've entered this in trac here: > > https://sourceforge.net/apps/trac/openvrml/ticket/92 I've been very short on time the last few months; so, unfortunately, it took me much longer to follow up on this than I would have preferred. However, I think this is finally fixed. The fix has been committed both to the trunk and to the 0.18 branch; I'd really appreciate it if you could test this in the next week. Now that this is finally fixed, I'm going to release 0.18.4 shortly. (Meanwhile I'll see if the Mozilla plug-in code can be doctored to compile with both XULRunner 1.9.1 *and* 1.9.2. Sigh. I'd really prefer not to break 1.9.1 in a bugfix release; OTOH, releasing something that doesn't work with 1.9.2 at this date would be similarly uncool.) -- Braden McDaniel <br...@en...> |