On Jan 31, 2011, at 12:34 AM, Peter Quirk wrote: > Toni - I filed this as incident 67 with crash dump yesterday.
Thanks! The guys at Evo get the PDBs out hopefully soon so we can see what it exactly it is. I put this info to the bug tracker too. I could finally reproduce it -- happened on my laptop when started to test the Opensim scene backup tool with which Cecil had probs, and boom! Usually use Linux on that thing, only tested this crash with a couple windows pcs at our studio earlier and couldn't reproduce on those. There is a simple workaround as a quick fix: disable UiExternalModule . This can be done by renaming Naali0.4.0/modules/core/UiExternalModule.xml to e.g. UiExternalModule.NOT (something that doesn't have xml in the name). I tested and without uiexternal buildmode works. Was good to get to see it, 'cause apparently it succeeds in switching to the new buildmode UI scene and in adding some of the widgets there (new object, clone object etc buttons) but then hits I think a null pointer before the parts with Inventory&Library and Components&Properties show. Earlier was guessing the prob would have been in the qt scene change but apparently that part works. The ideas of UiExternal and BuildMode are a bit conflicting, 'cause UiExt has generic automated docking to side panels, with tabbing, that work for any widgets. The older BuildMode makes similar panels but the config is hardcoded to certain widgets. Idea being that it's a handy preconfiguration of the widgets you typically need when authoring. Enne first used UiExternal without build mode at all, and instead made a 'object tools' menu entry that mimicked old buildmode by opening the widgets that builmode had .. as separate panels that you could then dock etc. if you wanted. That seemed like a good approach, but was not perfect at least because the extra 3d view to show the selected object was missing. So we figured to keep the old buildmode ui scene impl too for now at least. Unfortunately it was a bit tricky to keep working with how the ui widgets are managed now and apparently there is some bug that weirdly enough hits on some machines only. Luckily it seems simple so perhaps easy to fix once we get the debug info out, and at least there is that workaround for now. > Peter ~Toni > On Jan 28, 11:58 pm, Toni Alatalo <[email protected]> wrote: >> On Jan 29, 2011, at 6:21 AM, Peter Quirk wrote: >> >>> It produces a mini crash dump each time. The console error is "Program >>> realXtend encountered an unexpected error. Crash dump was saved to >>> location ...". >>> Where would you like me to place the crash dump? >> >> Oh great that it does, without the dump it is pretty impossible to know what >> is going on in the native code.. perhaps you can submit an issue to the >> google code tracker and attach the dump there? I think the only place now >> where we can debug it now is on the machine at Evocativi/Adminotech where >> the build for the release was made, but hopefully someone there can check >> the dump on Monday. And/or copy the debugging support files over so can do >> it elsewhere too. >> >>> Peter >> >> ~Toni >> >> >> >>> On Jan 28, 6:15 am, Toni Alatalo <[email protected]> wrote: >>>> On to, 2011-01-27 at 18:30 -0800, Peter Quirk wrote: >> >>>>> Did anyone find the fix? I'm having the same problem on two different >>>>> Windows 7 systems. I deleted appdata\Roaming\realXtend and c:\Users >>>>> \<user>\.naali-cookies, but it didn't change anything. >> >>>> I'm clueless so far. Just installed 0.4.0 on win7 and have no probs, >>>> build mode works etc. >> >>>> Do you get a crash dump? Or errors in the console? The console output >>>> goes to a log file, or alternatively you can run viewer.exe from command >>>> line when the output stays in the console after the crash. Probably >>>> there is no relevant output, though, but only a crash dump or being able >>>> to reproduce the problem when running with a debugger would help. >> >>>> I'll test on more machines. >> >>>>> - Peter >> >>>> ~Toni >> >>>>> On Jan 25, 10:04 am, Toni Alatalo <[email protected]> wrote: >>>>>> On Jan 25, 2011, at 2:47 PM, Peter Steinlechner wrote: >> >>>>>>> Yes, I tested the builds you mentioned and it sounds like its the same >>>>>>> issue you must have had. Thanks for this hint and I will try laters to >>>>>>> find those files. By any chance you have a hint about the file name ? >> >>>>>> 16:00 < Pforce> appdata/etc../configuration >>>>>> 16:00 < Pforce> .ini files from there >>>>>> 16:01 < Pforce> prolly some uiexternal.ini >>>>>> 16:01 < Pforce> matti had many crashes before clearing that file >> >>>>>> .. i think it's the appdata in your users home dir and realXtend dir >>>>>> there etc., am not on windows not so can't check >> >>>>>>> Pedro >> >>>>>> ~copy-paste-man >> >>>>>>> On Tue, Jan 25, 2011 at 1:25 PM, Toni Alatalo <[email protected]> wrote: >>>>>>> On Jan 24, 2011, at 5:20 PM, Peter Steinlechner wrote: >>>>>>>> When i try to use the build tool the viewer crashes no matter if I use >>>>>>>> it in Fisu or on the local Taiga 0.2.0 >> >>>>>>> Did you test the Enne builds at some point on that machine, the ones >>>>>>> where the external windows + docking feats were first introduced? I had >>>>>>> the prob on one box that got crashes with 0.4 builds 'cause had >>>>>>> incompatible .ini files (with the window setup saved) lying around. >> >>>>>>> if that's the case, I don't know yet what ini file / where to remove, >>>>>>> but we can figure out. >> >>>>>>>> Pedro >> >>>>>>> ~Toni >> >>>>>>>> On Mon, Jan 24, 2011 at 4:14 PM, Toni Alatalo <[email protected]> >>>>>>>> wrote: >>>>>>>> On Jan 21, 2011, at 1:45 PM, mattiku wrote: >>>>>>>>> Naali viewer 0.4.0 is now ready for use. >> >>>>>>>> Apparently a bug sneaked in that broke old mediaurls (the ones >>>>>>>> attached to textures with the old viewer, not using EC_3dCanvasSource >>>>>>>> directly) -- had happened in a refactoring I made a while ago. Just >>>>>>>> noticed it today 'cause some friendly person had added one to >>>>>>>> world.realxtend.org:9000 and the test bot that logs in there after >>>>>>>> each commit reported the fail. >> >>>>>>>> Fix is in develop now, and at least the bot doesn't complain anymore. >>>>>>>> Will test more at some point, and I guess we'll have to do a fix >>>>>>>> release soon'ish.. can wait a while if people report other possible >>>>>>>> probs too. >> >>>>>>>>> Matti Kuonanoja >> >>>>>>>> ~Toni >> >>>>>>>> -- >>>>>>>> http://groups.google.com/group/realxtend >>>>>>>> http://www.realxtend.org >> >>>>>>>> -- >>>>>>>> http://groups.google.com/group/realxtend >>>>>>>> http://www.realxtend.org >> >>>>>>> -- >>>>>>> http://groups.google.com/group/realxtend >>>>>>> http://www.realxtend.org >> >>>>>>> -- >>>>>>> http://groups.google.com/group/realxtend >>>>>>> http://www.realxtend.org-Hidequoted text - >> >>>>>> - Show quoted text -- Hide quoted text - >> >>>> - Show quoted text - >> >>> -- >>> http://groups.google.com/group/realxtend >>> http://www.realxtend.org- Hide quoted text - >> >> - Show quoted text - > > -- > http://groups.google.com/group/realxtend > http://www.realxtend.org -- http://groups.google.com/group/realxtend http://www.realxtend.org
