Re: GtkAssistant API thoughts
- From: Tristan Van Berkom <tristan van berkom gmail com>
- To: Matthias Clasen <mclasen redhat com>
- Cc: gtk-devel-list gnome org, Damon Chaplin <damon karuna uklinux net>
- Subject: Re: GtkAssistant API thoughts
- Date: Thu, 08 Sep 2005 14:48:47 -0400
Matthias Clasen wrote:
[...]
I understand that a user-specified function is a bit problematic for
a gui builder, but without the page objects, there will not be
per-page "forward" and "back" signals anyway, at most it would be
a forward signal on the assistant object, taking the current page or
page number as an argument. The main reason why I think a callback
function might be more convenient than a signal is that the assistant
code may want to call it repeatedly to determine the complete sequence
of "future" pages from the current page on (e.g. to determine whether
we should display a "Last" or "Finish" button). Maybe that is
possible with a signal as well...
To do that we need a way to call the "forward" signal/userfunc without
the page actually changing as a result, in that case we could use a "next-page"
delagate who returns the logical next page and a "display-page" to actually
do the "forward" / "backward" updates (both of which of course have a simple
default implementation), this scheme would also help to implement the quick
reference buttons to previously completed pages.
But sorry I'm drifting here; I was just thinking about the drawback of
having one signal handler for all pages when you only want to handle
a few pages... would it be too complex to use a signal detail like
"forward::page number" ? (other than that I dont think using a signal
changes much except the normal slight overheads).
Cheers,
-Tristan
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]