Melchior FRANZ
> Sent: 19 June 2007 12:23 > To: flightgear-devel@lists.sourceforge.net > Subject: Re: [Flightgear-devel] [ANN] OSG - Improved Weather Radar > > > * Vivian Meazza -- Sunday 17 June 2007: > > That's done - the patches are attached. The are NOT formatted > > properly, so no rants about tabs, spaces or trailing spaces. > > That's OK for the old code (and less so for the added code. > :-} Meanwhile the files for the base package are available, > too (even committed), so testing is actually possible. > > > The patches applied cleanly and compiled with a few warnings. > I found only minor things to fix: > > - dead but uncommented code in one case [if (foo && false) ...] > - redundant assignments [float x = 0; x = foo->getFloatValue();] > - compiler warnings > > I didn't thoroughly review (nor understand ;-) all the code, > especially not the OSG parts, but I trust Tim. Also, the > patches don't touch much other code, so I wouldn't be worried > about it. > > > The patch did not work for me at first, because (like other > developers, I guess) I'm using /sim/sceneryloaded-override. > This prevented that /sim/sceneryloaded ever became "true", > while od_gauge waited exactly for that. This is meanwhile > fixed (main.cxx). > > > After that the code worked in both fg/plib and fg/osg, but in > fg/plib I get a segfault on exit, which comes from > RenderTexture.cpp. That's quite a hairy piece of code, and > I'm not really competent to fix it. I checked on the net if > newer RenderTexture implementations have that code part > fixed, but this is not the case. > > > #0 0x7773612f in ?? () > #1 0xb7c4f52f in _X11TransWritev () from > /usr/lib/libX11.so.6 #2 0xb7c54f21 in _XSend () from > /usr/lib/libX11.so.6 #3 0xb7c4625b in XQueryExtension () > from /usr/lib/libX11.so.6 #4 0xb7c3ab0b in XInitExtension () > from /usr/lib/libX11.so.6 #5 0xb6ffb0f3 in XextAddDisplay () > from /usr/lib/libXext.so.6 #6 0xb7db368e in > glXChannelRectSyncSGIX () from /usr/lib/libGL.so.1 #7 > 0x08a69258 in ?? () #8 0x08b008c0 in ?? () #9 0xb7de2eb5 in > std::basic_streambuf<char, std::char_traits<char> > >::showmanyc () from /usr/lib/libGL.so.1 #10 0xb7df2dc0 in > std::basic_streambuf<char, std::char_traits<char> > >::showmanyc () from /usr/lib/libGL.so.1 #11 0x00000011 in ?? > () #12 0x08b008c0 in ?? () #13 0x1137ead8 in ?? () #14 > 0x08589295 in RenderTexture::_Invalidate (this=0xb7db5fe0) at > simgear/screen/RenderTexture.cpp:848 > #15 0x0858ea8f in ~RenderTexture (this=0x1137ead8) at > simgear/screen/RenderTexture.cpp:204 > #16 0x083b9e22 in ~FGODGauge (this=0xb45e518) at > src/Instrumentation/od_gauge.cxx:89 > #17 0x085d6b2d in ~Member (this=0xd1ea7a0) at > simgear/structure/subsystem_mgr.cxx:227 > #18 0x085d7ba9 in ~SGSubsystemGroup (this=0xb7dadb67) at > simgear/structure/subsystem_mgr.cxx:85 > > > So, applying the patches for fg/plib would mean to replace a > cheesy but not-crashing radar implementation by a nice but > crashing one. I don't say that the radar patch is buggy, it's > just the old render-to-texture feature. (It's also not my > graphics card driver, as Qt4 has no problems with RTT.) > > I would very much like to apply the patches, but I think the > crash should be fixed first. (Or should the fg/osg patches go > in, anyway?) > That crash is not repeatable here, but perhaps that wouldn't be unexpected. I am surprised by it though - the code (od_gauge.cxx) which interacts with RenderTexture.cpp is untouched by this patch. Perhaps this isn't a new bug after all. Perhaps someone else could confirm this behaviour? Vivian ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
