Bug 99449 – WebPageProxy::canShowMIMEType() should return false for plugins when they are disabled

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

[Carlos Garcia Campos]

It currently returns true, but if the web view tries to handle the plugin it ends up rendering garbage.

[Carlos Garcia Campos]

Created attachment 168915 [details]
Patch

[Rik Cabanier]

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

[Alexey Proskuryakov]

Could you please post a URL of a web page that's affected? Since this changes cross-platform code, I'd like to see this issue in practice on Mac.

If not handled, what will happen with plug-in resources? Will they be downloaded?

[Xan Lopez]

(In reply to comment #3)
> Could you please post a URL of a web page that's affected? Since this changes cross-platform code, I'd like to see this issue in practice on Mac.

Not sure what page was Carlos using to test this. I have been using http://www.ikea.com/fr/fr/, which contains a SWF resource.

> 
> If not handled, what will happen with plug-in resources? Will they be downloaded?

That is mostly up to the port and/or the embedder. In WebKitGTK+/Epiphany we'll download the resource if it's the main resource (so that clicking on link that points to a PDF file will download it), but ignore anything else (so that a <object> tag pointing to a SWF file will be ignored if Flash is not supported).

[Alexey Proskuryakov]

There is indeed a garbage subframe in Safari on this IKEA page.

For Apple employees: see also <rdar://problem/11650197>.

[Carlos Garcia Campos]

(In reply to comment #5)
> There is indeed a garbage subframe in Safari on this IKEA page.

Yes, this is exactly what happens in MiniBrowser if plugins are disabled.

> For Apple employees: see also <rdar://problem/11650197>.

[Anders Carlsson]

Comment on attachment 168915 [details]
Patch

I don't think the shouldBlockPlugin check is correct here. We should try (and fail) to load the plug-in when it's blocked. Please remove that check.

[Carlos Garcia Campos]

Created attachment 169356 [details]
Updated patch

[Anders Carlsson]

Comment on attachment 169356 [details]
Updated patch

r=me

[Carlos Garcia Campos]

Committed r131867: <http://trac.webkit.org/changeset/131867>