Re: Why does my cursor doesn't change

[Emmanuele Bassi <ebassi gmail com>]

hi Aristotle;

On Wed, 2007-04-04 at 17:51 +0200, A. Pagaltzis wrote:
> * Emmanuele Bassi <ebassi gmail com> [2007-04-03 18:50]:
> > On Tue, 2007-04-03 at 11:36 -0500, Kenneth Swanson wrote:
> > > Are we not telling people the other, quick 'n' dirty way to
> > > get around this?  That is, that if you have a long block of
> > > code in a callback or some other code, you can add the line
> > >
> > > Gtk2->main_iteration while(Gtk2->events_pending);
> > >
> > > to get the GUI to update?  Or does that not work correctly
> > > anymore?
> >
> > it's not the "kosher" solution: forcing the main loop to spin
> > it's an hack, and masks the real problem
>
> Thatâs true, but does it have any *tangible* drawbacks to kick
> the mainloop manually?

not that I know of, but it's really not nice apart from a stop-gap
solution while you fix your code.

once you add it and you keep relying on it, you usually end up
replicating that line time and again around critical sections and the
readability and maintainability of your code begins to suffer.

>  Is there any reason why this shouldnât be
> used even in cases where it would be *much* harder to punt the
> processing to a timeout callback (or a forked child)?

using an idle or a timeout source never is *much* harder: it's not
different than using a while(1){} loop, not even conceptually.

spinning the main loop manually is the solution you apply when
porting/fixing code that has been written by someone else more versed in
console applications, when you need to phase out that code while people
are still working with it.

ciao,
 Emmanuele.

-- 
Emmanuele Bassi,  E: ebassi gmail com
W: http://www.emmanuelebassi.net
B: http://log.emmanuelebassi.net