[webkit-gtk] Architectural doubt

[Martin Robinson]

On Thu, Feb 10, 2011 at 2:02 AM, Sergio Villar Senin <svillar at igalia.com> wrote:
> in order to implement some missing DRT features I found that I needed
> access to some LayoutTestController calls already implemented in the
> header file, from the FrameLoaderClientGtk. That's not a big deal on mac
> port for for example, as their Delegates have direct access to the
> LayoutTestController API, but that's not the case in our port. What do
> you think is the "right way" to do that?

I had a similar problem with the editing delegate callbacks the first
time I tried to implement them. I wanted to just put the print
statements in the EditorClient in WebKit, but the setting which
flipped them on and off only existed passively in
LayoutTestController. The Mac port just checked it before every print
message.

The solution for this was to implement them in the "right way," by
which I mean actually plumb the signals through and expose them
publically. They are undocumented, but they are there now.

--Martin