Bug 74605 – [GTK] Simplify loader client WebKit2 GTK+ API

• Component: WebKit2
• Status: RESOLVED FIXED

[Carlos Garcia Campos]

The initial idea of the loader client object exposed in the API was mainly to reduce the amount of signals of WebView. During the webkitgtk hackfest we were discussing the current API and agreed on moving to simpler API adding just two signals to WebView and making the current loader client private.

[Carlos Garcia Campos]

Created attachment 119424 [details]
Patch

The code is simpler and the API easier to use. Now we also have only one condition for load finish. The patch doesn't include previous status in load-status-changed callback as agreed during the hackfest, because Martin has some doubts about it, it's better to add it later if we finally decide it's useful than remve it.

[WebKit Review Bot]

Thanks for the patch. If this patch contains new public API please make sure it follows the guidelines for new WebKit2 GTK+ API. See http://trac.webkit.org/wiki/WebKitGTK/AddingNewWebKit2API

[Martin Robinson]

Comment on attachment 119424 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=119424&action=review

> Source/WebKit2/UIProcess/API/gtk/WebKitWebView.cpp:316
> +     * @status: the #WebKitLoadStatus

I prefer @event: the #WebKitLoadEvent here. See below.

> Source/WebKit2/UIProcess/API/gtk/WebKitWebView.h:71
> +} WebKitLoadStatus;

So I think this should be renamed to  WebKitLoadEvent since WEBKIT_LOAD_REDIRECTED is not really a status, but something that happened to us.

[Carlos Garcia Campos]

(In reply to comment #3)
> > Source/WebKit2/UIProcess/API/gtk/WebKitWebView.h:71
> > +} WebKitLoadStatus;
> 
> So I think this should be renamed to  WebKitLoadEvent since WEBKIT_LOAD_REDIRECTED is not really a status, but something that happened to us.

hmm, good point!

[Martin Robinson]

(In reply to comment #4)
> (In reply to comment #3)
> > > Source/WebKit2/UIProcess/API/gtk/WebKitWebView.h:71
> > > +} WebKitLoadStatus;
> > 
> > So I think this should be renamed to  WebKitLoadEvent since WEBKIT_LOAD_REDIRECTED is not really a status, but something that happened to us.
> 
> hmm, good point!

I guess we can go two ways here: try to make the enum represent a status or an events For instance if we used events, they enum values would be something like:

WEBKIT_LOAD_STARTED_PROVISIONAL
WEBKIT_LOAD_REDIRECTED
WEBKIT_LOAD_COMMITTED
WEBKIT_LOAD_FINISHED

and the argument names would be something like previous_load_event instead of previous_load_status.

If we think of them as states we would have:

WEBKIT_LOAD_PROVISIONAL
WEBKIT_LOAD_REDIRECTING
WEBKIT_LOAD_COMMITTED
WEBKIT_LOAD_FINISHED

Personally, I prefer the 'event' structure because that's the interface that the C API gives us and I think it more closely matches what's happening within WebCore.

[Gustavo Noronha (kov)]

That's a great approach, Martin, I agree with you!

[Carlos Garcia Campos]

Created attachment 120538 [details]
Updated patch

Renamed WebKitLoadStatus to WebKitLoadEvent. I also renamed load-status-changed signal as load-changed to avoid the word 'status' and for consistency with load-failed.

[Carlos Garcia Campos]

Created attachment 120539 [details]
I forgot to include Tools/ChangeLog in previous patch

[Gustavo Noronha (kov)]

Comment on attachment 120539 [details]
I forgot to include Tools/ChangeLog in previous patch

View in context: https://bugs.webkit.org/attachment.cgi?id=120539&action=review

OK, looks good to me

> Source/WebKit2/UIProcess/API/gtk/WebKitWebView.h:57
> + * fail for transport issues such as not being able to

s/for/due to/ sounds better I think

[Carlos Garcia Campos]

Committed r103938: <http://trac.webkit.org/changeset/103938>