[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: port of pyjamas to pykde / khtml - issues and observations with
From:       "Luke Kenneth Casson Leighton" <lkcl () lkcl ! net>
Date:       2008-10-13 17:33:47
Message-ID: ced5f0f60810131033j411fdf56x2ef5379cdbff3f94 () mail ! gmail ! com
[Download RAW message or body]

> how much better would it be to have the KHTML code - and pykde - do
> the work for every single developer, already?

also, crucially - absolutely, absolutely crucially: you can't do a
"compare" against nodes.  the objects are potentially different.  and
that has _massive_ implications, not least of which is that if you put
member variables into a python object which is a wrapper around a c++
object, you're not guaranteed to get the same python object _back_...

this is royally screwing up the pyjamas-khtml port.  it was fixed very
easily in webkit-glib, such that it was a complete non-issue in
pywebkitgtk.  the key was in the use of a virtual base class for the
"wrapper" class, which, for every single type of object returns a
GOBJECT type.  the GDOMBinding.cpp module is then responsible for
allocating the _correct_ type of binding object to represent the c++
object and for placing into the HashMap.

l.
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]