Peter Hutterer said the following on 10/17/2012 12:12 AM: > On Tue, Oct 16, 2012 at 04:40:14PM +0200, Olivier Fourdan wrote: >> Hi Peter, >> >> Peter Hutterer said the following on 10/16/2012 07:35 AM: >>> [...] >>> git bisect claims this broke make check (dbverify segfaults). Can you please >>> double-check that? thanks. >> Well, actually, this is not a segfault but an assert() which fails. >> >> Reason it fails is because it compares the definition in the old database >> against the new one, problem these patches add new information that were not >> present in the older database so these tests will necessarily fail each time >> we add new information to the existing entries of the database. > dbverify doesn't really have a old/new database though. it loads the > database once, writes it to a file, then re-reads that file as new database > and compares it to the original. Theory goes that the two databases have to > be identical, otherwise either write or read is broken. > > In my case it breaks on the first I4 tablet, so I suspect the LED stuff > either doesn't print correctly, compare correctly or read correctly*. This is > the real problem we need to find.
Oh crap yes, it's not the right patch that was pushed in git :-( The one which was pushed in git reads "StatusLEDs=" and writes "LEDs=" so it would indeed fail... The correct one was: http://sourceforge.net/mailarchive/message.php?msg_id=29918039 I shall send a fix to get to the same level as that one then. Cheers, Olivier. ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_sfd2d_oct _______________________________________________ Linuxwacom-devel mailing list Linuxwacom-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel