Re: XForms: Mysterious crashing bug is 0.88

Steve Lamont (spl@szechuan.ucsd.edu)
Sat, 7 Feb 98 17:19:59 PST

# To subscribers of the xforms list from spl@szechuan.ucsd.edu (Steve Lamont) :

> The problem occurs on windows which have a single input object,
> along with buttons, pixmaps, and inactive free objects. The form
> is controlled through a small main loop that calls fl_check_forms()
> along with some other status checking. Some of the pixmaps and free
> objects are being updated (via fl_redraw_object()) either from the
> main loop or via timeout callbacks which poll external devices.

Admittedly, this is a wild guess but are you doing anything in the
Free object handlers to manipulate any other XForms objects? As I
mentioned in another recent posting, doing this can screw up XForms's
internal bookkeeping and could, I suppose, cause crashes.

Another thing to try might be to turn your main loop inside out and
have the status checking routine called as a idle callback.

A couple of things you might try:

Does this happen in the `fdesign' test mode, as well? Does it happen
if you simply create the user interface without the supporting code?
You could try this with probably only a small amount of effort by
copying the .fd file to another directory, running `fdesign' to create
the user interface, the callbacks, and a main routine, then
compiling. You can use the `fdesign' Options menu to generate almost
all of this.

I did see this sort of behavior, or something quite similar, in 0.87
-- in fact, it was so serious and pervasive I had to fall back to 0.86
-- but I haven't seen it in 0.88 at all.

spl
_________________________________________________
To unsubscribe, send the message "unsubscribe" to
xforms-request@bob.usuf2.usuhs.mil or see
http://bob.usuf2.usuhs.mil/mailserv/xforms.html
XForms Home Page: http://bloch.phys.uwm.edu/xforms
List Archive: http://bob.usuf2.usuhs.mil/mailserv/list-archives/