Re: window.c abort() crash in 21.4.6-14
Darryl Okahata <darrylo(a)soco.agilent.com> writes:
gak(a)klanderman.net (Greg Klanderman) wrote:
> I've been seeing xemacs abort in the following code in window.c:
That code is there, in theory, to catch any "infinite loop" problems
in window_loop().
Do you have a lot of XEmacs "windows" (i.e., frames split multiple
times into multiple "XEmacs windows", not actual X11 windows)? I'm
wondering if the "500" value is too low, when lots of XEmacs
"windows" are present.
I don't see how the value 500 could be too low for anything resembling
normal usage. It's far more realistic to assume that we're still
seeing the infinite loop bug this used to guard against. :-(
[ For that matter, is it possible to have non-visible XEmacs
windows?
I believe windows deleted and re-created from the window configuration
code, but the non-visible windows should not be reachable through the
buffer attributes. If they are, it's just another form of the bug.