Re: [Flightgear-devel] crashes in GL
Vivian Meazza wrote: And SG? Before you tear your system apart, AJ has just reported similar symptoms over on IRC. He's just updated to CVS HEAD. Yes, both FG and SG from 7th July. -- Jon Stockill [EMAIL PROTECTED] ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
RE: [Flightgear-devel] crashes in GL
Jon Stockill > > I rolled back cvs to 7th July, built flightgear, and still have exactly > the same problem - I guess it's related to something on this system. > > -- > Jon Stockill > [EMAIL PROTECTED] > And SG? Before you tear your system apart, AJ has just reported similar symptoms over on IRC. He's just updated to CVS HEAD. Vivian ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] crashes in GL
I rolled back cvs to 7th July, built flightgear, and still have exactly the same problem - I guess it's related to something on this system. -- Jon Stockill [EMAIL PROTECTED] ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] crashes in GL
Vivian Meazza wrote: No, but if it's any compensation, cvs has been broken under Cygwin since, well, I've tracked it back to around 7th Jul. For me it started when I downloaded a clean copy. Norman Vine's version is exiting as soon as the load sequence finishes, mine and a machine AJ is testing on hangs at the same point. At the moment the problem seems to be in Simgear, possibly in Harald's new GL stuff, but that's a very tentative diagnosis. Right now I can't see anything obviously wrong, and am really at a loss as to what to do next. The last known good seems to be 7th Jul - you might like to roll cvs (FG and SG) back to that point and see what happens. I'd be interested to compare notes I'll give that a try. -- Jon Stockill [EMAIL PROTECTED] ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
RE: [Flightgear-devel] crashes in GL
Jon Stockill > > Harald JOHNSEN wrote: > > Jon Stockill wrote: > > > >> Due to the death of the machine I was doing all my flightgear work on > >> I'm currently trying to set up another machine so I can still build > >> packages, but I've run into a bit of a problem. While my old packages > >> work ok on this machine (it's not gonna win any awards for framerates, > >> but it proves it works) the current cvs code segfaults on startup like > >> this: > >> > >> Program received signal SIGSEGV, Segmentation fault. > >> [Switching to Thread 16384 (LWP 18753)] > >> 0x401861b2 in _glthread_SetTSD () from /usr/X11R6/lib/libGL.so.1 > >> (gdb) bt > >> #0 0x401861b2 in _glthread_SetTSD () from /usr/X11R6/lib/libGL.so.1 > >> #1 0x401863f3 in glXCreateGLXPbufferSGIX () from > >> /usr/X11R6/lib/libGL.so.1 > > > > > > > >> #2 0x0846a17a in RenderTexture::Initialize (this=0x891d220, width=1, > >> height=1075494657, shareObjects=false, copyContext=false) > >> at RenderTexture.cpp:508 > > > > > > You are not supposed to have this width and height, in the bbcache.cxx > > code there is > > rt->Initialize(256, 256, true); > > This gets stranger and stranger. I rebuilt the code from scratch > (suggested by Harald), and it still failed with the same error. So I > deleted the whole thing, and checked it all out from CVS again - a > completely clean start. Same error. Checked out the code on another > machine, built it, runs ok. Copied the binary from the working machine > to the problem machine, and now the error is back - same binary, > different machines, different result - a segfault caused by an out of > range value which is supposed to be hard coded. > > Tests so far have ruled out plib/simgear/flightgear version problems, > and compiler problems (since the working binary fails on the problem > system). > > Does anyone have any idea where I should be looking before I go > completely insane? > No, but if it's any compensation, cvs has been broken under Cygwin since, well, I've tracked it back to around 7th Jul. For me it started when I downloaded a clean copy. Norman Vine's version is exiting as soon as the load sequence finishes, mine and a machine AJ is testing on hangs at the same point. At the moment the problem seems to be in Simgear, possibly in Harald's new GL stuff, but that's a very tentative diagnosis. Right now I can't see anything obviously wrong, and am really at a loss as to what to do next. The last known good seems to be 7th Jul - you might like to roll cvs (FG and SG) back to that point and see what happens. I'd be interested to compare notes Vivian ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] crashes in GL
Harald JOHNSEN wrote: Jon Stockill wrote: Due to the death of the machine I was doing all my flightgear work on I'm currently trying to set up another machine so I can still build packages, but I've run into a bit of a problem. While my old packages work ok on this machine (it's not gonna win any awards for framerates, but it proves it works) the current cvs code segfaults on startup like this: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 16384 (LWP 18753)] 0x401861b2 in _glthread_SetTSD () from /usr/X11R6/lib/libGL.so.1 (gdb) bt #0 0x401861b2 in _glthread_SetTSD () from /usr/X11R6/lib/libGL.so.1 #1 0x401863f3 in glXCreateGLXPbufferSGIX () from /usr/X11R6/lib/libGL.so.1 #2 0x0846a17a in RenderTexture::Initialize (this=0x891d220, width=1, height=1075494657, shareObjects=false, copyContext=false) at RenderTexture.cpp:508 You are not supposed to have this width and height, in the bbcache.cxx code there is rt->Initialize(256, 256, true); This gets stranger and stranger. I rebuilt the code from scratch (suggested by Harald), and it still failed with the same error. So I deleted the whole thing, and checked it all out from CVS again - a completely clean start. Same error. Checked out the code on another machine, built it, runs ok. Copied the binary from the working machine to the problem machine, and now the error is back - same binary, different machines, different result - a segfault caused by an out of range value which is supposed to be hard coded. Tests so far have ruled out plib/simgear/flightgear version problems, and compiler problems (since the working binary fails on the problem system). Does anyone have any idea where I should be looking before I go completely insane? -- Jon Stockill [EMAIL PROTECTED] ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] crashes in GL
Jon Stockill wrote: Due to the death of the machine I was doing all my flightgear work on I'm currently trying to set up another machine so I can still build packages, but I've run into a bit of a problem. While my old packages work ok on this machine (it's not gonna win any awards for framerates, but it proves it works) the current cvs code segfaults on startup like this: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 16384 (LWP 18753)] 0x401861b2 in _glthread_SetTSD () from /usr/X11R6/lib/libGL.so.1 (gdb) bt #0 0x401861b2 in _glthread_SetTSD () from /usr/X11R6/lib/libGL.so.1 #1 0x401863f3 in glXCreateGLXPbufferSGIX () from /usr/X11R6/lib/libGL.so.1 #2 0x0846a17a in RenderTexture::Initialize (this=0x891d220, width=1, height=1075494657, shareObjects=false, copyContext=false) at RenderTexture.cpp:508 You are not supposed to have this width and height, in the bbcache.cxx code there is rt->Initialize(256, 256, true); Perhaps you can set a break point on the call in bbcache and then follow the execution step by step to check what happens. Harald. ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] crashes in GL
Due to the death of the machine I was doing all my flightgear work on I'm currently trying to set up another machine so I can still build packages, but I've run into a bit of a problem. While my old packages work ok on this machine (it's not gonna win any awards for framerates, but it proves it works) the current cvs code segfaults on startup like this: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 16384 (LWP 18753)] 0x401861b2 in _glthread_SetTSD () from /usr/X11R6/lib/libGL.so.1 (gdb) bt #0 0x401861b2 in _glthread_SetTSD () from /usr/X11R6/lib/libGL.so.1 #1 0x401863f3 in glXCreateGLXPbufferSGIX () from /usr/X11R6/lib/libGL.so.1 #2 0x0846a17a in RenderTexture::Initialize (this=0x891d220, width=1, height=1075494657, shareObjects=false, copyContext=false) at RenderTexture.cpp:508 #3 0x08414179 in SGBbCache::init (this=0xa2cc490, cacheCount=64) at bbcache.cxx:125 #4 0x0840b604 in SGNewCloud::init (this=0xa239b1c) at newcloud.cxx:78 #5 0x0840b762 in SGNewCloud (this=0xa239ae0, classification=CLFamilly_cu) at newcloud.cxx:85 #6 0x08409691 in SGCloudField::buildTestLayer (this=0xa6db588) at cloudfield.cxx:297 #7 0x08402d9e in SGCloudLayer::rebuild (this=0xa762ce0) at cloud.cxx:468 #8 0x084024f6 in SGCloudLayer (this=0xa762ce0, [EMAIL PROTECTED]) at cloud.cxx:227 #9 0x08053f2c in fgIdleFunction () at sg_path.hxx:122 #10 0x0808a472 in GLUTidle () at fg_os.cxx:114 #11 0x400a1929 in __glutRegisterEventParser () from /usr/X11R6/lib/libglut.so.3 #12 0x400a20ad in glutMainLoop () from /usr/X11R6/lib/libglut.so.3 #13 0x080553d8 in fgMainInit (argc=1, argv=0xb794) at main.cxx:1011 #14 0x08051766 in main (argc=1, argv=0xb794) at bootstrap.cxx:193 I'm guessing it's a problem with the GL capabilities of this machine, and that we're now stressing th GL stuff more than we were before, but a segfault isn't a particularly graceful exit, so I thought I'd better let people know - maybe it's something we need to trap. Googling for _glthread_SetTSD throws up lots of problems and very few solutions. -- Jon Stockill [EMAIL PROTECTED] ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
