Thread: [Vimprobable-users] [Ticket #6] Handling new window links generated by Javascript

SourceForge Headquarters 1320 Columbia Street Suite 310 San Diego, CA 92101 +1 (858) 422-6466

Hi,

the attached patch should finally make the browser handle window.open events  
correctly. Testing of different websites would be greatly appreciated.

Hannes

On Mi, Jun 27 2012, Hannes Schüller wrote:
> Hi,
>
> the attached patch should finally make the browser handle window.open
> events correctly. Testing of different websites would be greatly
> appreciated.

I haven't tested this much, but so far it seems to work. I have
published a git branch with the patch for easier testing:

http://git.cs.fau.de/?p=lu03pevi/vimprobable;a=shortlog;h=refs/heads/fixes/javascript_windows

> -static gboolean
> -webview_open_in_new_window_cb(WebKitWebView *webview, WebKitWebFrame *frame, gpointer user_data) {
> -    Arg a = { .i = TargetNew, .s = (char*)webkit_web_view_get_uri(webview) };
> -    if (strlen(rememberedURI) > 0) {
> -        a.s = rememberedURI;
> -    }
> +void
> +webview_open_js_window_cb(WebKitWebView* web_view, GParamSpec param_spec) {
> +    Arg a = { .i = TargetNew, .s = (char*)webkit_web_view_get_uri(web_view) };
>      open_arg(&a);
> -    return FALSE;
> +}
> +
> +static WebKitWebView *
> +webview_open_in_new_window_cb(WebKitWebView *webview, WebKitWebFrame *frame, gpointer user_data) {
> +    WebKitWebView *new_view = WEBKIT_WEB_VIEW(webkit_web_view_new());

I haven't read the documentation for this, but doesn't "new_view" have
to be destroyed somewhere? Otherwise you'll be leaking memory.

> +    g_object_connect(new_view, "signal::notify::uri", G_CALLBACK(webview_open_js_window_cb), NULL, NULL);
> +    return new_view;
>  }

HP

On 28/06/12 at 02:35am, Hans-Peter Deifel wrote:
> On Mi, Jun 27 2012, Hannes Schüller wrote:
> > Hi,
> >
> > the attached patch should finally make the browser handle window.open
> > events correctly. Testing of different websites would be greatly
> > appreciated.
> 
> I haven't tested this much, but so far it seems to work. I have
> published a git branch with the patch for easier testing:
> 
> http://git.cs.fau.de/?p=lu03pevi/vimprobable;a=shortlog;h=refs/heads/fixes/javascript_windows
> 
Thanks this works great. 

There is slightly aberrant behaviour when opening a link from within Google
Reader (I know, no-one cares, but I wanted to mention it in case it pops up
anywhere else): hitting 'v' opens the linked article in a new window and also
opens duckduckgo (default search engine) with an 'about:blank' search.

/J

-- 

http://jasonwryan.com/  [GnuPG Key: B1BD4E40]

Jason Ryan <jas...@gm...> wrote:
> On 28/06/12 at 02:35am, Hans-Peter Deifel wrote:
> > On Mi, Jun 27 2012, Hannes Sch__ller wrote:
> > > the attached patch should finally make the browser handle
> > > window.open events correctly. Testing of different websites would
> > > be greatly appreciated.
> > 
> > I haven't tested this much, but so far it seems to work. I have
> > published a git branch with the patch for easier testing:
> > 
> > http://git.cs.fau.de/?p=lu03pevi/vimprobable;a=shortlog;h=refs/heads/fixes/javascript_windows
> > 
> Thanks this works great. 
> 
> There is slightly aberrant behaviour when opening a link from within
> Google Reader (I know, no-one cares, but I wanted to mention it in
> case it pops up anywhere else): hitting 'v' opens the linked article
> in a new window and also opens duckduckgo (default search engine)
> with an 'about:blank' search.

Who says no-one cares? Please show me an example (which I don't have to
sign up for). If I can reproduce it, I will try to fix this. Though, of
course, I'd rather have two windows opened than none ;)

Hannes

On 28/06/12 at 10:32am, Hannes Schüller wrote:
> Jason Ryan <jas...@gm...> wrote:
> > On 28/06/12 at 02:35am, Hans-Peter Deifel wrote:
> > > On Mi, Jun 27 2012, Hannes Sch__ller wrote:
> > > > the attached patch should finally make the browser handle
> > > > window.open events correctly. Testing of different websites would
> > > > be greatly appreciated.
> > > 
> > > I haven't tested this much, but so far it seems to work. I have
> > > published a git branch with the patch for easier testing:
> > > 
> > > http://git.cs.fau.de/?p=lu03pevi/vimprobable;a=shortlog;h=refs/heads/fixes/javascript_windows
> > > 
> > Thanks this works great. 
> > 
> > There is slightly aberrant behaviour when opening a link from within
> > Google Reader (I know, no-one cares, but I wanted to mention it in
> > case it pops up anywhere else): hitting 'v' opens the linked article
> > in a new window and also opens duckduckgo (default search engine)
> > with an 'about:blank' search.
> 
> Who says no-one cares? Please show me an example (which I don't have to
> sign up for). If I can reproduce it, I will try to fix this. Though, of
> course, I'd rather have two windows opened than none ;)
> 
I meant that as it was a Google-related site, I would not expect much interest
from the Vimprobable community (and rightly so…)

I've only seen it on that site and it does require a sign on; and yes, you are
right, it works twice as well as it used to!

/J
-- 

http://jasonwryan.com/  [GnuPG Key: B1BD4E40]

Hans-Peter Deifel writes:
> I haven't read the documentation for this, but doesn't "new_view" have
> to be destroyed somewhere? Otherwise you'll be leaking memory.

Here is a revised version of the patch (also with decent comments).

Hannes

Hannes Schüller writes:
> Hans-Peter Deifel writes:
>> I haven't read the documentation for this, but doesn't "new_view" have
>> to be destroyed somewhere? Otherwise you'll be leaking memory.
>
> Here is a revised version of the patch (also with decent comments).

This version has been merged.

Hannes

Thread: [Vimprobable-users] [Ticket #6] Handling new window links generated by Javascript

Vimprobable is a lean web browser optimised for full keyboard control

vimprobable-users