[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

[Vala] Lets talk about some easier stuff before this controversial one.



On Fri, 2008-08-22 at 11:54 +0200, Jürg Billeter wrote:
> On Fri, 2008-08-22 at 05:24 -0400, Yu Feng wrote:
> > On Fri, 2008-08-22 at 11:04 +0200, Jürg Billeter wrote:
> > > > Another direction of this patch is to allow small wrapping code in
> > > > the .vapi declarations (which might ease the vapi authors's life)
> > > 
> > > This is the only place where I might consider it useful. However, so far
> > > I'm not convinced that it really helps, see my other response.
> > > 
> > intended for binding (again, using some asm wrap a call to int13) and I
> > began with .vala files because I thought they were easier to do
> > experiment with. 
> > 
> > Perhaps the binding authors can find more examples? in Samba and GST
> > bindings? If we put it to upstream people will eventually find use
> > cases. It should be rare.
> > 
> > To pull the vala GList wrappers out I need to patch the <generic>
> > (valagenis?) to allow overriding with respect of generic. The patch is
> > already large and it is not wise to make it larger before positive
> > responses.
> > 
> > How much can this vapi-only patch hurt vala in a long run?
> 
> The patch is already getting quite big, and it seems that you've added a
> partial C parser in there. 
That's what I was worrying about several month ago. The parser parses
identifiers and look for the formal parameters, the result, and "this". 
I thought it was difficult but it turned out not much. 

> Before I'm considering maintaining this, I'd
> need a strong use case, so far I haven't seen one that would justify it.
> 
I am happy as long as I've made it possible. I understand your feelings
too. Therefore I won't be very unhappy if the patch was hung there.
Nevertheless we can try to separate useful stuff(if there is) from the
patch first.
See
http://bugzilla.gnome.org/show_bug.cgi?id=548969

If this works we can then remove
---
    public override string get_default_cname () {
        // return cname of wrapper method
        if (cname == null) {
            cname = "_dynamic_%s%d".printf (name, dynamic_method_id++);
        }
        return cname;
    }
---
in valadynamicmethod.vala
and kill the replica wrappers.

> Jürg
> 



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]