[Webkit-unassigned] [Bug 16401] [GTK] GObject/C DOM binding

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=16401





------- Comment #199 from lkcl at lkcl.net  2009-06-07 01:53 PDT -------
(In reply to comment #86)
> (In reply to comment #52)
> > --- WebCore/page/DOMWindow.idl  (revision 36459)
> > +++ WebCore/page/DOMWindow.idl  (working copy)
> > @@ -155,7 +155,7 @@
> >          [DoNotCheckDomainSecurity, Custom] void postMessage(in DOMString
> > message, in DOMString targetOrigin)
> >              raises(DOMException);
> > 
> > -#if defined(LANGUAGE_JAVASCRIPT)
> > +#if defined(LANGUAGE_JAVASCRIPT) || defined(LANGUAGE_GOBJECT)
> >          // Global constructors
> >          attribute StyleSheetConstructor StyleSheet;
> >          attribute CSSStyleSheetConstructor CSSStyleSheet;
> > 
> > i've added this in as a reminder that, at some point _in the future_,
> > the gobject auto-generator code will need to auto-generate "constructors".
> > i believe this was part of the problem i had in doing creation of a
> > document - not having some of these functions.
> > 
> > although it will be straightforward to add constructors, i'm not doing
> > it for this patch.
> 
> Constructors are JavaScript language concept.  They're not something that other
> bindings need to worry about.

 unfortunately, as has just been discovered, not having constructors,
 and making them "javascript language concept", means that certain objects
 and certain functionality can never be accessed.

 there is an implicit assumption that all functionality - all objects -
 can be accessed somehow as the return result of another function, where,
 absolute best case, you can have one and only one function at the top-level
 such as gdom_get_dom_window() and then access quite literally everything
 from there.

 unfortunately, that is not the case.

 "XPathNSResolver" is one object which is NOT returned as the return
 result from ANY other webkit DOM object function.

 in amongst the 450 or so webkit DOM objects, there will be several others.

 this is the primary reason why constructors were added to javascript,
 and you would do well to stop interfering with the development of the
 glib / gobject language bindings and start trusting the experience and
 process which created the javascript language bindings over several
 years.

 there are _good_ reasons why the javascript language bindings are the
 way that they are, and it's going to be _really_ fucking boring to have
 to justify at every single step of the way every single fucking feature
 again and again.

 a much simpler process is if you (apple) stop interfering and blocking
 development by trying to dictate what these glib / gobject language
 bindings _should_ have, according to some idealistic W3C standard instead
 of the de-facto real world javascript standard, and start trusting the
 judgement of the people who are actually _using_ these language bindings
 in the real world.


-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.