Re: Subclassing Gtk2::CellRenderer

[muppet <scott asofyet org>]

On Thursday, October 2, 2003, at 07:43 AM, Bjarne Steinsbø wrote:

> Still, wouldn't it be nice to be able to do anything you can do in C 
> in perl as well?  Not necessarily making it simple to do in perl, but 
> at least possible?

yes, it would be nice to have all the C functionality in perl, but it 
comes at a price, and for some things the price is just too high.


> Some additional magic during code generation, maybe?  Exposing in 
> special-purpose perl classes what you're doing your best to hide in 
> the normal case?

i spent a very long time several months ago trying to figure this one 
out, and a good while last night looking over it again to see if it'd 
missed anything.  (i had a very long section here about how i'd go 
about implementing generic overridable methods, to illustrate why we 
don't do it, but in the end it was just to long and muddled.  if you 
really want to know, ask later and i'll explain.)  that whole idea got 
left by the wayside when i finally discovered that 
g_signal_override_class_closure was the sanctioned way to do this, even 
though it has limitations.

i think, though, it would be possible to add some special stuff to the 
bindings for just this one class (GtkCellRendererClass) to see what we 
can do to add support for overriding those functions.

but be warned, this is "new function" stuff, so please don't expect it 
to be part of the stable release.  (it'll take some bribery otherwise.  
;-)


> An optionmenu shouldn't really be that hard to render in the normal 
> case, it's just a string and a pixmap.  It's the editing phase that's 
> complicated.  But I see your point.

from what i can tell, you just need something like 
gtk_paint_drop_shadow around the edges of the box and maybe some sort 
of indicator that you can click on this thing to pop up a menu.  the 
string will be easy, using pango rendering.  the fun comes in when you 
have to calculate  box dimensions and such.  i'm interested in seeing 
the difference in performance between perl and C.


i have to run... contact me offlist if you want to try something like 
this.


> P.S. I think I stumbled across a typo in the xs documentation for 
> Glib.  Please see attached diff.

eek!  thanks!

--
muppet <scott at asofyet dot org>