Excerpts from Ben Gamari's message of Tue Oct 06 12:15:46 -0400 2009: > Excerpts from Ben Gamari's message of Tue Oct 06 11:53:18 -0400 2009: > > Well, it seems like whatever caused the crash earlier did something to > > my index. Now any attempt to open a thread-index-mode of my LKML label > > (which I was viewing in the earlier crash) causes the client to > > immediately crash. > > > > Hmm, it seems like the problem is spreading. I now come to find out that > another label triggers this same crash (although I guess it's possible > that the intersection of the two labels is non-nil). I tried running a > sup-sync -oc on the relevant sources to no avail. I really don't have > time to devote to debugging this at the moment so it looks like I might > need to take another hiatus from sup. Just as I was starting to get used > to it... sigh. Anyways, if anyone has any ideas for improving things, > let me know. Thanks!
I've been seeing this crash for a long time. I think it's a race between the poll thread / load-more-threads thread and the rest of the UI in the main thread. Basically any operation on a thread object in ThreadIndexMode needs to be protected against ThreadSet#add_message (and probably other ThreadSet methods) because add_message can remove containers from the thread tree, leaving you with an empty thread whose "date" method returns nil. You could try running sup with the -n flag to disable threading. The major downside is that you have to hit P to poll manually. I look forward to having a sup-server - I plan on writing a little android client in Scala using actors. Almost no mutable state and absolutely no ncurses. _______________________________________________ sup-talk mailing list [email protected] http://rubyforge.org/mailman/listinfo/sup-talk
