Highly reproducible, but machine-specific: on a Thinkpad E530c: Behaviour: 1) open Writer, new document, and ctrl-enter to add three or four pages -- or else: open any document with more than a few pages. 2) scroll down, then up with the touchpad 3) X freezes; sometimes freezing completely; sometimes hanging the GPU. I have two machines, both runnning up-to-date Debian Sid; the Thinkpad E530c does this, making Writer difficult to use; the Thinkpad T420i has no problems. This happens in both LO 4.0.1 and 3.5.4.2, the version in the Debian sid repositories. I filed this as a bug against the debian xserver-xorg-video-intel package some time ago (for logs, see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=703749 ), but: there's been no movement; and this only occurs with Libreoffice, so I am posting it here to find out how to tell if the bug is with LO or with X somehow. Please let me know how to diagnose this. thanks, --p
petrelharp: could you try to retrieve a backtrace? (see https://wiki.documentfoundation.org/BugReport#How_to_get_a_backtrace_on_Linux)
Created attachment 77532 [details] backtrace of the crash The SIGTERM is because I had to ssh in (frozen X) and kill soffice to get this. Let me know if there's more info needed.
petrelharp: thank you for your feed back, it might be useful indeed! http://opengrok.libreoffice.org/xref/core/vcl/unx/generic/app/wmadaptor.cxx#2194 2206 if( XGetWindowProperty( m_pDisplay, 2207 aWindow, 2208 m_aWMAtoms[ NET_WM_DESKTOP ], 2209 0, 1, 2210 False, 2211 XA_CARDINAL, 2212 &aRealType, 2213 &nFormat, 2214 &nItems, 2215 &nBytesLeft, 2216 &pProperty ) == 0 2217 && pProperty 2218 ) 2219 { 2220 nCurrent = int(*(sal_Int32*)pProperty); 2221 XFree( pProperty ); 2222 } 2223 else if( pProperty ) 2224 { 2225 XFree( pProperty ); 2226 pProperty = NULL; 2227 } I noticed a double cast line 2220, I don't know if it's normal or not, perhaps it is. Maybe pProperty should be put to NULL also in "if" block? Caolán: one for you? Put it at New since we got a bt.
Unfortunately this appears to be a bug on the other side, and there doesn't seem to be much we can do about it.
Can you tell where the bug is -- is it the intel drivers, or elsewhere? Any other information that might be useful? thanks a bunch, p.