Since r116356 seamless frames support is optional and disabled in GTK. So the following tests fail: Regressions: Unexpected text diff mismatch : (10) fast/frames/seamless/seamless-form-get-named.html = TEXT fast/frames/seamless/seamless-form-get.html = TEXT fast/frames/seamless/seamless-form-post-named.html = TEXT fast/frames/seamless/seamless-form-post.html = TEXT fast/frames/seamless/seamless-hyperlink-named.html = TEXT fast/frames/seamless/seamless-hyperlink.html = TEXT fast/frames/seamless/seamless-window-location-href.html = TEXT fast/frames/seamless/seamless-window-location-sandbox.html = TEXT fast/frames/seamless/seamless-window-location.html = TEXT fast/frames/seamless/seamless-window-open.html = TEXT I'll flag those in test_expectations for now but I think we should simply enable the feature in GTK. Any thoughts, Martin?
Adam, do you think it's safe to start shipping this feature in unstable releases and then in a stable release in about 6 months?
> Adam, do you think it's safe to start shipping this feature in unstable releases and then in a stable release in about 6 months? Yes. Eric is still landing some pieces of the feature, but six months is plenty of time to bake.
Okay. I'd say it's quite safe to turn this feature on. it carries a lot less risk than some of our API developer facing features and we'll be able to test it more thoroughly by turning it on now.
Created attachment 141195 [details] Patch
(In reply to comment #4) > Created an attachment (id=141195) [details] > Patch This patch adds a configuration option in configure.ac to enable this feature, but is disabled by default. The feature is enabled by default in build-webkit, though. I can however enable this feature as well in configure.ac as the comment #3 hints.
Comment on attachment 141195 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=141195&action=review > Tools/Scripts/webkitperl/FeatureList.pm:196 > + define => "ENABLE_ICONDATABASE", default => isGtk(), value => \$iframeSeamlessSupport }, Looks like the define is wrong here. It should probably be ENABLE_IFRAME_SEAMLESS.
(In reply to comment #6) > (From update of attachment 141195 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=141195&action=review > > > Tools/Scripts/webkitperl/FeatureList.pm:196 > > + define => "ENABLE_ICONDATABASE", default => isGtk(), value => \$iframeSeamlessSupport }, > > Looks like the define is wrong here. It should probably be ENABLE_IFRAME_SEAMLESS. Caught red-handed :> A proper patch incoming.
Created attachment 141209 [details] Patch
Comment on attachment 141209 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=141209&action=review > Tools/Scripts/webkitperl/FeatureList.pm:196 > + define => "ENABLE_IFRAME_SEAMLESS", default => isGtk(), value => \$iframeSeamlessSupport }, should still be default => 1 > configure.ac:742 > + [],[enable_iframe_seamless="no"]) I stillt hink this should defatult to yes.
Comment on attachment 141209 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=141209&action=review >> Tools/Scripts/webkitperl/FeatureList.pm:196 >> + define => "ENABLE_IFRAME_SEAMLESS", default => isGtk(), value => \$iframeSeamlessSupport }, > > should still be default => 1 Given that Mac port showed some desire for an option to ship with this feature disabled, I guess this shouldn't affect them? >> configure.ac:742 >> + [],[enable_iframe_seamless="no"]) > > I stillt hink this should defatult to yes. Figures.
Created attachment 141213 [details] Patch
Comment on attachment 141213 [details] Patch build-webkit option defaults are wildly inaccurate until I fix bug 85456 anyway. :) If you're interested in this topic, we still need someone to split the Gtk feature defaults out into some easily autogenerated file to fix that bug!
Comment on attachment 141213 [details] Patch Clearing flags on attachment: 141213 Committed r116679: <http://trac.webkit.org/changeset/116679>
All reviewed patches have been landed. Closing bug.