Re: [gtk-list] Qt succeeds where gtk fails... (callbacks)

[robert_gasch peoplesoft com]

Hi Eric,

while I'm no expert on timeout/callback internals, I have managed
to write Gtk-- (the C++ wrapper) code which sucessfully uses a
timeout ... after reading your message I started wondering wether
it worked properly (I've probably never let it run for more than a
minute before). I turned it on and let it run and after 1/2 hour it's still
chugging away, so it seems to work as it should ... I'm currently
using Gtk-- 1.0.1, Gtk+1.0.3, Glib-1.0.3 on a RH 6.0 system ...

Greetings
--> Robert





Eric Livingston <elivings@clark.net> on 07/02/99 04:05:23 PM

Please respond to gtk-list@redhat.com

To:   gtk-list@redhat.com
cc:
Subject:  [gtk-list] Qt succeeds where gtk fails... (callbacks)


GTK seems to develop amnesia on me and "forget" that programs such as
gpppkill and gstripchart have one-second timeout callback functions that
they use to update their displays. They both start fine, but after several
seconds, they stop updating. Inserting some printfs here and there has led
me to realize that after a short time, gtk simply stops calling the
functions, even if the timeout functions consistenly return non-zero
values (thus, they should not be unregistered).

I just downloaded and am testing a Qt program called pppload, which does
similar things to gpppkill, and which I can also set to update every
second.

That program has been working fine now for over an hour.

I'm beginning to think very strongly that there is a fundamental flaw in
how gtk is handling its callbacks...

Can anyone guide me in trying to debug this further? I've already
determined that the select() function seems to be working fine on my box,
so it's got to be somewhere else in gtk where this problem manifests...

Eric Livingston


--
To unsubscribe: mail -s unsubscribe gtk-list-request@redhat.com < /dev/null