Re: notebook api a pain...

["Michael K. Johnson" <johnsonm redhat com>]

Tim Janik writes:
>On Mon, 31 Aug 1998, Michael K. Johnson wrote:
>> The notebook API is a pain for certain uses.  In particular, if you
>> want to keep track of a page you have added, you have to jump through
>> hoops.  For example, I have an application where I need to add pages
>> in no particular order, then remove them at random.  There is no real
>> way to identify a page right now.  :-(
>
>not fully correct. you add a page by providing the (container) widget
>that needs to appear as a notebook page later on. this widget pointer
>is more or less the handle you'd like to have.

Um, what did I say that was incorrect?

Certainly, that widget pointer could be used as the handle -- that is
fine -- but what is important to me is that I *have* a handle, whether
it is that pointer or not.  I want a handle of some sort that DOES NOT
change when arbitrary pages are removed.  You have suggested one
reasonable way to implement what I suggest.  That doesn't mean that
what I wrote was incorrect; what I wrote (that there is no current
facility for this) is perfectly correct.  The only way I can tell to
do it is to happen to know that it is maintained internally as a
linked list and figure it all out magically, which sucks.

>there needs to be a function, though, that returns the current
>page number for a given notebook child, i.e.
>
>gint gtk_notebook_get_page_num (GtkNotebook *notebook,
>                                GtkWidget   *page);
>
>which will return -1 if page is not a child of the notebook, and its
>position otherwise.

Yes, that would work well.

michaelkjohnson

"Magazines all too frequently lead to books and should be regarded by the
 prudent as the heavy petting of literature."            -- Fran Lebowitz
 Linux Application Development       http://www.redhat.com/~johnsonm/lad/