Bug 38135 – Include RuntimeEnabledFeatures in the Qt build

• Component: WebCore Misc.
• Status: RESOLVED INVALID

[Garret Kelly]

Currently, RuntimeEnabledFeatures is not included in the Qt build. This patch adds it to the Qt build, alongside a modification in WebSockets to remove the guards that stop runtime-feature support from only being enabled under the V8 platform.

[Garret Kelly]

Created attachment 54325 [details]
Initial patch

[Kenneth Rohde Christiansen]

Comment on attachment 54325 [details]
Initial patch

What is this exactly? How do we enable this? Do we need to add a setting to the QWebSettings class?

If you can answer those questions, I will happily review the patch.

[Eric Seidel (no email)]

Why does Qt want this?

[Csaba Osztrogonác]

(In reply to comment #3)
> Why does Qt want this?

I'm not convinced if Qt port would like to use it, but this patch blocks https://bugs.webkit.org/show_bug.cgi?id=37485

[David Levin]

Comment on attachment 54325 [details]
Initial patch

WebCore/websockets/WebSocket.cpp:81
 +  static bool webSocketsAvailable = true;
There is no explanation for this change in the ChangeLog.


WebCore/ChangeLog:5
 +          Include RuntimeEnabledFeatures in the Qt build.
It sounds like folks who work on Qt aren't convinced that this is a change they want.


Why not just have a implementation of RuntimeEnabledFeatures that just returns true and use that for Qt (and include it with the other patch that needs it)?

[Kenneth Rohde Christiansen]

(In reply to comment #5)

> It sounds like folks who work on Qt aren't convinced that this is a change they
> want.

I'm not really against it, but I just have no idea what it is :-) thus it is very hard to give an opinion on whether this is a change we want or not.

[Csaba Osztrogonác]

As far as I see only Chromium port uses RuntimeEnabledFeature now. I think it can be useful for testing experimental features on other ports. If we really want to use this feature into Qt port, it's OK. But I think the goal of this patch is to "Make touch event support a RuntimeEnabledFeature" ( https://bugs.webkit.org/show_bug.cgi?id=37485 ) for the Chromium port without breaking Qt port. If I am correct, it isn't the good solution, but papering over the cracks.

[Garret Kelly]

Papier-mâché seems like a poor approach. I've changed the proposed patch for 37485 and am closing this issue.