From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040312
Description of problem:
It takes about 1 minute to crash the vncviewer when working on a
Windows machine running VNC 4.0 Beta 3. After the viewer crashes you
can try to reconnect to the server and it allows you to authenticate,
but as soon as it trys to display the desktop again it crashes.
GDB backtrace is included.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. I connect to our Windows 2000 Server running VNC 4.0 Beta 3
2. Open a few windows, look at some logs
3. vncview crashes unexpectedly
4. crashes instantly after connecting back up again.
Expected Results: Not crashing, the windows VNC Viewer works just
fine against the same server.
Tue May 4 13:40:37 2004
CConn: connected to host 172.16.10.10 port 5900
CConnection: Server supports RFB protocol version 3.7
CConnection: Using RFB protocol version 3.7
Tue May 4 13:40:40 2004
TXImage: Using default colormap and visual, TrueColor, depth 24.
CConn: Using pixel format depth 8 (8bpp) colour-map
CConn: Using ZRLE encoding
Program received signal SIGSEGV, Segmentation fault.
0x08067d51 in TXImage::lookup (this=0x90a1088, index=-66424728,
g=0x90a1088, b=0xfc0a7068) at TXImage.cxx:166
166 *r = colourMap[index].r;
#0 0x08067d51 in TXImage::lookup (this=0x90a1088, index=-66424728,
r=0x90a1088, g=0x90a1088, b=0xfc0a7068) at TXImage.cxx:166
#1 0x08074476 in rfb::PixelFormat::rgbFromPixel (this=0x90a1090,
p=4276794024, cm=0x90a10c0, rgb=0xfeeab2b0) at PixelFormat.cxx:111
#2 0x0804cefa in DesktopWindow::setCursor (this=0x90a0b00,
hotspot=@0xfc0a7068, size=@0x90a10c0, data=0xfc0a7068, mask=0x220005f)
#3 0x080852cd in rfb::CMsgReader::readSetCursor (this=0x9099330,
hotspot=@0xfc0a7068, size=@0xfeeab3a8) at CMsgReader.cxx:144
#4 0x0806bda5 in rfb::CMsgReaderV3::readMsg (this=0x9099330) at Rect.h:39
#5 0x0806a77b in rfb::CConnection::processMsg (this=0x90a1088)
#6 0x080634bb in main (argc=1, argv=0xfeeab440) at vncviewer.cxx:314
#7 0x002aef43 in __libc_start_main () from /lib/tls/libc.so.6
#8 0x0804bfe1 in _start ()
Does upgrading the Windows side to 4.0beta4 make the problem go away?
There were several cursor-handling problems fixed since beta3.
The windows server is already running 4.0b4 built Sept 4th 2003, that
was my mistake in the bug report.
Can you capture the traffic of a complete session, starting from the
viewer connecting, and finished after it segfaults, using tcpdump please?
Something like: tcpdump -w capture.pcap port 5901
Then attach the capture.pcap file.
Never got around to closing this bug, resending the message I meant to
Upgrading the server on the windows side to 4.0beta5 (and final) seems
to have resolved it. If I can reproduce it with this version I'll
reopen the bug report.