Re: GtkButtons and RUN_FIRST/RUN_LAST

[Owen Taylor <otaylor redhat com>]

Tim Janik <timj gtk org> writes:

> On 26 Oct 2001, Havoc Pennington wrote:
> 
> > 
> > Manish Singh <yosh gimp org> writes: 
> > > At this point, everything works for me with current GTK+, so I'm not
> > > really as concerned as I was before, but something still doesn't sit right
> > > about the states not being resolved in a ::clicked user handler.
> > 
> > It makes perfect sense though. Clicked notifies that the button was
> > clicked. In response to the click, toggle button toggles its state,
> > thus emitting toggled. But the toggle is in response to the click, not
> > in anticipation of the click.
> > 
> > Ah well, backward compat is a bastard.
> 
> this is not just a pure backwards compat issue, so that "we're screwed
> because things were wrong the first time".
> the original behaviour was actually intended and part of our API
> contract, though, it might not be how you had things designed.

Luckily for progress, settling this is pretty much irrelevant at this
point - the fact that Joel and Manish and you are doing bad hacks with
connections to ::clicked that anticipate what the default handler did
indicates that changing this at this point would be too incompatible,
no matter what is correct.

> can you please put gtkbutton back to RUN_FIRST now?

I'll do this. Someday, I'll write my own toolkit, and such things will
not be tolerated, goshdarnit! ;-(

                                        Owen