Bug 82795 – Add tests for iframe seamless and support for parsing webkitseamless attribute

• Status: RESOLVED
• Resolution: LATER
• Component: New Bugs

[Eric Seidel (no email)]

Add tests for iframe seamless and support for parsing webkitseamless attribute

[Eric Seidel (no email)]

Created attachment 134904 [details]
Patch

[Adam Barth]

Comment on attachment 134904 [details]
Patch

Do we need to announce that we're implementing seamless to webkit-dev ?

[Eric Seidel (no email)]

(In reply to comment #2)
> (From update of attachment 134904 [details])
> Do we need to announce that we're implementing seamless to webkit-dev ?

Done: https://lists.webkit.org/pipermail/webkit-dev/2012-March/020158.html

[Ojan Vafai]

YAY!

[WebKit Review Bot]

Comment on attachment 134904 [details]
Patch

Clearing flags on attachment: 134904

Committed r112760: <http://trac.webkit.org/changeset/112760>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Alexey Proskuryakov]

I've seen it multiple times now that people first land failing tests, and then start implementing a feature. Where is this idea coming from?

I think that this is exceptionally unhelpful and wrong. The reviewer looking at actual patch won't see what test coverage we have. Having ad hoc tests for unimplemented features is just a weird state for the code base to be in.

>  // FIXME: webkit prefix will be removed when implementation is complete.

So why is this not behind an ifdef? A port that branches for release now won't want to expose this implementation artifact to the web.

[Adam Barth]

> I've seen it multiple times now that people first land failing tests, and then start implementing a feature. Where is this idea coming from?

I can't speak for Eric, but I often write patches this way.  It's advocated by some folks under the name test-driven development:

http://en.wikipedia.org/wiki/Test-driven_development

> >  // FIXME: webkit prefix will be removed when implementation is complete.
> 
> So why is this not behind an ifdef? A port that branches for release now won't want to expose this implementation artifact to the web.

There's a discussion about this point on webkit-dev.  If you're interested in this topic, I encourage you to share your thought there.

[Alexey Proskuryakov]

> It's advocated by some folks under the name test-driven development:

Writing tests first is what I strongly advocate, too. Landing these is an entirely different matter.

[Eric Seidel (no email)]

Reverted r112760 for reason:

Revert addition of webkitseamless.  I'll do this work on GitHub instead to avoid any half-implemented feature concerns.

Committed r112820: <http://trac.webkit.org/changeset/112820>

[Eric Seidel (no email)]

The GitHub branch is here:
https://github.com/eseidel/webkit/compare/master...seamless