[Glade-devel] GladePalette patch

[Tristan Van Berkom]

On Sat, 18 Dec 2004 20:16:31 -0500, Archit Baweja
<bighead@users.sourceforge.net> wrote:
> Hey ho!
> 
> So, here's another patch. It basically changes the GladePalette's
> catalog selection implementation. Instead of using individual radio
> buttons, I used a GtkComboBox. The reason behind this decision is that
> since glade3 has extensible catalog support, we can have more than 4
> catalogs in the future. And having individual radio button for each
> would ... well would just not be right.
> 
> Also, I took the oppurtunity to cleanup the code a bit. Here I would
> suggest a little programming tip . Whenever using
> g_object_[get/set]_data() functions, for the paramenter name's string,
> don't use a string, but use a all caps constant (or a #define, smae
> thing). Because if you mis-type it or change it, the compiler will catch
> it for you. Plus, if someone decides to change the string he/she must
> find all the places in the code where to change and that can be tedious.

Hi, 
      GtkComboBox IMO is a good idea (it will save the user significant screen
realestate), we might want to converge to a glade-2 style palette for the
same reason. Code-cleanup is also welcome ofcourse (there is still much 
cleanup to be done). 

About the "Cancel" button in the query popup, implementation looks
sound to me, although I didn't look at the patch in depth yet.

About using the GladeCommand interface inside the generic Query popup, 
yes that is definitly a bug but I think the "from_query_dialog" code should
stay in glade-editor.c, if there is not need for an undo; then we should
skip the middle-man and jump directly to glade_property_set().

Finally, please send patches to bugzilla; I find it alot easier to
track that way (nothing gets lost that way and I can always identify
which is the latest version of your patch).

I'll review the patch[es] monday or maybe tomorow, things might get
a little rocky this week since I'll be integrating the libglade branch,
so If your patch gets applied first; I might ask you to review the merge 
afterwards ;-)

Cheers,
                                                         -Tristan