On Thu, 11 Jan 2007, Tom Russo wrote: > Just began to look into those segfaults tonight. I can reliably crash > xastir (built with lesstif) by repeatedly opening the send-message dialog > and changing the format between d700/hamhud/d7, then closing the dialog. > > The segfault is in a free, down several layers of motif calls.
I can still get LSB-Xastir, which is compiled with Lesstif, to segfault as well, so there's more work to do on the Send Message dialog before this is put to rest for LSB. Once that work is complete we could test for Lesstif in the code and implement the LSB fixes in that case. > Perhaps it would be best not to try the on-the-fly changing of the dialog, > since clearly lesstif has some serious Issues with that? How many dialogs > do things this way as opposed to the close-dialog, change, redisplay method? Problem is, that's part of the Motif definition. They specifically say that some apps might even start up incomplete and keep adding widgets before your eyes. The "manager" widgets are supposed to handle changing the geometry to account for that. Motif handles this beautifully. I wouldn't mind munging the code if Lesstif is detected, but I _really_ like how the widgets changed based on HamHUD/D700/D7 detection. I find this quite useful and would like to keep it for any Motif compiles. I don't _think_ we currently do this sort of thing anyplace else in the code, but I recall trying to do it before and having difficulties with Lesstif back then. That's why some of the changes are implemented by destroying and recreating widgets from scratch, like the Create Object dialog. I find that sort of operation annoying though: The dialog flashes at that point plus gets placed in a new spot when it comes back up. -- Curt, WE7U. APRS Client Comparisons: http://www.eskimo.com/~archer "Lotto: A tax on people who are bad at math." -- unknown "Windows: Microsoft's tax on computer illiterates." -- WE7U "The world DOES revolve around me: I picked the coordinate system!" _______________________________________________ Xastir mailing list [email protected] http://lists.xastir.org/cgi-bin/mailman/listinfo/xastir
