Bug 42546 – WebKitTestRunner needs layoutTestController.setCustomPolicyDelegate

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKit2

[Maciej Stachowiak]

WebKitTestRunner needs layoutTestController.setCustomPolicyDelegate

[Maciej Stachowiak]

<rdar://problem/8204916>

[Balazs Kelemen]

We have created a Qt specific bug for this issue. Later we found out that it is not platform specific and removed the [Qt] prefix. I mark this as duplicate because #42330 already has patches uploaded. Sorry for the confusion. If it is important to you to track the issue here please invert the direction of duplication and we will reupload the patch.

*** This bug has been marked as a duplicate of bug 42330 ***

[Balazs Kelemen]

Sorry for the noise. I missed it, these are not the same issue. :-S

[Fehér Zsolt]

I have a patch, that look like this: https://gist.github.com/1156343
One test failed because of an unknown reason.
http/tests/security/feed-urls-from-remote.html
The test has a beginTests() method. But this method didn't run. Do somebody have any idea, why not?

[Balazs Kelemen]

Created attachment 111958 [details]
Patch

[Adam Roben (:aroben)]

Comment on attachment 111958 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=111958&action=review

> Source/WebKit2/WebProcess/InjectedBundle/API/c/WKBundlePage.h:137
>  enum {
> +    WKBundlePagePolicyActionIgnore,
>      WKBundlePagePolicyActionPassThrough,
>      WKBundlePagePolicyActionUse

I believe this breaks ABI compatibility with Safari 5.1 (i.e., will break nightly builds).

> Tools/ChangeLog:17
> +        waitToPolicyDelegate. We need to ignore the navigation action if we
> +        are not in permissive mode. This has not caused a problem with
> +        waitToPolicyDelegate because with that we explicitly stop testing
> +        via notifyDone and do not send any request after that.

What is "waitToPolicyDelegate"?

> LayoutTests/platform/wk2/Skipped:564
>  # WebKitTestRunner needs layoutTestController.setCustomPolicyDelegate
>  # <https://bugs.webkit.org/show_bug.cgi?id=42546>
> -fast/loader/javascript-url-hierarchical-execution.html
> -fast/loader/onload-policy-ignore-for-frame.html
> -fast/loader/reload-policy-delegate.html
> -http/tests/misc/policy-delegate-called-twice.html
> -http/tests/misc/redirect-to-external-url.html
> -http/tests/security/feed-urls-from-remote.html
> +#fast/loader/javascript-url-hierarchical-execution.html
> +#fast/loader/onload-policy-ignore-for-frame.html
> +#fast/loader/reload-policy-delegate.html
> +#http/tests/misc/policy-delegate-called-twice.html
> +#http/tests/misc/redirect-to-external-url.html
> +#http/tests/security/feed-urls-from-remote.html

You should delete the lines rather than just commenting them out.

[Balazs Kelemen]

Created attachment 112308 [details]
Patch

[Balazs Kelemen]

(In reply to comment #6)
> (From update of attachment 111958 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=111958&action=review
> 
> > Source/WebKit2/WebProcess/InjectedBundle/API/c/WKBundlePage.h:137
> >  enum {
> > +    WKBundlePagePolicyActionIgnore,
> >      WKBundlePagePolicyActionPassThrough,
> >      WKBundlePagePolicyActionUse
> 
> I believe this breaks ABI compatibility with Safari 5.1 (i.e., will break nightly builds).

Moved the new one to the end.

> 
> > Tools/ChangeLog:17
> > +        waitToPolicyDelegate. We need to ignore the navigation action if we
> > +        are not in permissive mode. This has not caused a problem with
> > +        waitToPolicyDelegate because with that we explicitly stop testing
> > +        via notifyDone and do not send any request after that.
> 
> What is "waitToPolicyDelegate"?

Added a quick explanation.

> 
> > LayoutTests/platform/wk2/Skipped:564
> >  # WebKitTestRunner needs layoutTestController.setCustomPolicyDelegate
> >  # <https://bugs.webkit.org/show_bug.cgi?id=42546>
> > -fast/loader/javascript-url-hierarchical-execution.html
> > -fast/loader/onload-policy-ignore-for-frame.html
> > -fast/loader/reload-policy-delegate.html
> > -http/tests/misc/policy-delegate-called-twice.html
> > -http/tests/misc/redirect-to-external-url.html
> > -http/tests/security/feed-urls-from-remote.html
> > +#fast/loader/javascript-url-hierarchical-execution.html
> > +#fast/loader/onload-policy-ignore-for-frame.html
> > +#fast/loader/reload-policy-delegate.html
> > +#http/tests/misc/policy-delegate-called-twice.html
> > +#http/tests/misc/redirect-to-external-url.html
> > +#http/tests/security/feed-urls-from-remote.html
> 
> You should delete the lines rather than just commenting them out.

Fixed.

[Balazs Kelemen]

Comment on attachment 112308 [details]
Patch

It's not the desired solution according to IRC discussion.
The UI process should know about what's going on.

[Mikhail Pozdnyakov]

Could I take it (since there was no activity on this bug for almost a year)?

[Balazs Kelemen]

(In reply to comment #10)
> Could I take it (since there was no activity on this bug for almost a year)?

Sure, feel free to take it.

[Mikhail Pozdnyakov]

(In reply to comment #11)
> (In reply to comment #10)
> > Could I take it (since there was no activity on this bug for almost a year)?
> 
> Sure, feel free to take it.
Thanks!

[Mikhail Pozdnyakov]

Created attachment 162772 [details]
patch

[Mikhail Pozdnyakov]

(In reply to comment #13)
> Created an attachment (id=162772) [details]
> patch

Did not ask cq? as this patch should land after bug95974 is resolved so that fast/loader/onload-policy-ignore-for-frame.html passes.

[Mikhail Pozdnyakov]

Created attachment 164118 [details]
to be landed

Skipped fast/loader/onload-policy-ignore-for-frame.html to remove dependency to bug95974, as looks like it will take more time to land..

[WebKit Review Bot]

Comment on attachment 164118 [details]
to be landed

Clearing flags on attachment: 164118

Committed r128600: <http://trac.webkit.org/changeset/128600>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Balazs Kelemen]

Are we really ready with this? It still not goes through the UI process.