[Webkit-unassigned] [Bug 44261] [GTK] Add HTTP caching support

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=44261





--- Comment #3 from Sergio Villar Senin <svillar at igalia.com>  2010-08-20 01:56:48 PST ---
(In reply to comment #2)
> (From update of attachment 64849 [details])
> This patch is huge! I'd really like to see it split up into at least
> two parts.
> 
> 1. The code drop from Soup, which should probably be separated somehow from
> the rest of our source, so we can treat it specially (it will go away eventually, right?).
> 2. The bits that actually implement the cache in WebCore/WebKit.

Thx for reviewing Martin.

Well basically the patch consists of:
   1) changes in ResourceHandleSoup/ResourceHandleInternal: basically adaptations to work with the URILoader stuff
   2) WebCore/platform/network/soup/webkitsoup*: all the stuff moved from webkit
   3) WebKit/Wekit/gtk/webkit/webkitsoupcache.[c|h]: the webkitgtk+ API for clients

Now some comments:

> > +    if (m_req) {
> > +        g_object_set_data(G_OBJECT(m_req.get()), "webkit-resource", 0);
> > +        m_req.clear();
> 
> There's no chance that this may be the last reference to m_req here?

What's the problem then?

> > +    GOwnPtr<char> uri(soup_uri_to_string(soup_message_get_uri(msg), false));
> > +    String location = String(uri.get());
> 
> Even though URL are generally URL-encoded, I don't know if we can rely on that
> here. You should use String::fromUTF8.

soup_uri_to_string always returns an URL encoded URL, do you think we still need to use ::fromUTF8 ?

> > -    d->m_idleHandler = g_timeout_add(0, parseDataUrl, handle);
> > +    d->m_idleHandler = g_idle_add(parseDataUrl, handle);
> 
> Why did you switch this to an idle? I think Gustavo is trying to move
> away from that because they are often starved. Maybe ping him about this.

As stated here https://bugs.webkit.org/show_bug.cgi?id=40222#c11 there was 1 test that fails with timeout but not with idle. After a long time debugging it, it seems that it's just a matter of timing. All the signals and events are emitted properly but somehow in a different order depending on the function used.

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