Bug 15561 – GTK port needs DumpRenderTree implementation

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Eric Seidel (no email)]

GTK port needs DumpRenderTree implementation

See attached patch.

[Eric Seidel (no email)]

Created attachment 16729 [details]
partial patch

[Holger Freyther]

Add the gtk keyword.

[Holger Freyther]

*** Bug 14735 has been marked as a duplicate of this bug. ***

[Holger Freyther]

Add the blocking as well...

[Eric Seidel (no email)]

Created attachment 16738 [details]
updated partial patch

[Eric Seidel (no email)]

Created attachment 16743 [details]
a more complete patch

Ok, I give up.  Someone else will have to finish this.

[Eric Seidel (no email)]

Created attachment 16744 [details]
a more complete patch (previous one was missing files)

[Alp Toker]

Created attachment 17456 [details]
WebKitFrame changes needed to support DRT

This patch implements and documents more of the WebKitFrame public API.

[Maciej Stachowiak]

Comment on attachment 17456 [details]
WebKitFrame changes needed to support DRT

I'm ok with contributing to this file under LGPL assuming other WebKit/Gtk hackers are ok with this direction. However, I don't think you can just delete the BSD-style license. I think you need to preserve the notice, although of course the LGPL will take precedence.

The code changes look ok to me but I think I should r- over the license detail. Please fix and resubmit.

[Alp Toker]

(In reply to comment #9)
> (From update of attachment 17456 [details] [edit])
> I'm ok with contributing to this file under LGPL assuming other WebKit/Gtk
> hackers are ok with this direction. However, I don't think you can just delete
> the BSD-style license. I think you need to preserve the notice, although of
> course the LGPL will take precedence.

The only way you can re-license is to completely remove the old text, otherwise new code will remain optionally available under the BSD license, right?

[Alp Toker]

Created attachment 17470 [details]
WebKitFrame changes needed to support DRT, with fixed licensing

Guess we'll have to punt on the re-licensing until those issues are solved.

[Mark Rowe (bdash)]

Comment on attachment 17470 [details]
WebKitFrame changes needed to support DRT, with fixed licensing

r=me, except you should be using 0 consistently instead of both 0 and NULL.

[Alp Toker]

Comment on attachment 17470 [details]
WebKitFrame changes needed to support DRT, with fixed licensing

Patch already landed, but the bug report needs to remain open. Removing r+ flag to clear the commit queue.

[Xan Lopez]

Created attachment 17641 [details]
0001-First-stub-at-DRT-for-Gtk.patch

Patch updated to work with trunk.

Few things I had to do (also using WorkQueueItemGtk from Alp):

- Implement get_children and get_inner_text for WebKitWebFrame.
- Make JSStringCopyUTF8CString return something.
- Make ScriptItem::invoke() use script() instead of url()
- Generally update to use new APIs

And probably some other things I forget :)

Have to run now, please have a look at the patch and later I can update a new version addressing any comment and with ChangeLog etc.
-

[Xan Lopez]

Another small comment: the Qt port seems to have a class explicitely named FrameLoaderClientQt, while we have two classes named FrameLoaderClient, but in different namespaces. Is this intentional?

[Alp Toker]

Comment on attachment 17641 [details]
0001-First-stub-at-DRT-for-Gtk.patch

r=me

This patch needs some amount of style cleanup, less clobbering of the shared headers, and the new API needs to be made internal-only till discussed, but basically fine and a good start.

[Alp Toker]

Landed in r28384. Still needs lots of work.

I'll watch the build bots to make sure nothing breaks, and disable DRT in the default GTK+ build if it causes trouble.

Thanks Xan!