May 2012 - webkit-efl - lists.webkit.org

Dropping support for the pango network backend.
by ryuan Choi 17 Aug '12

17 Aug '12

Hello, Efl folks. We've supported freetype and pango backend as font backend of webkit-efl. However, pango backend can not render well because of missing features and webkit-efl builds with freetype by default. Does anyone need pango backend? If possible, I will remove font backend selection checks and just use freetype backend with FontPango.cpp for the complex font. Related bug is https://bugs.webkit.org/show_bug.cgi?id=86369 Best Regards, Ryuan Choi

4 8

Building WebKit2?
by Dominik Röttsches 01 Jun '12

01 Jun '12

Hi all, I am trying to build the WebKit2 flavour of EFL WebKit using: $ Tools/Scripts/build-webkit --efl --debug --cmakearg="-D^CARED_CORE=ON -DENABLE_WEBKIT2=ON" Is that the right approach? Anything else I should be aware of? It currently fails with a CMAKE error that I'm investigating. If it is the right way I would go and document it on the Wiki. Dominik -- Dominik Röttsches<dominik.rottsches(a)intel.com>

5 6

ewk_view settings implementation
by Pozdnyakov, Mikhail 31 May '12

31 May '12

Hello, I am implementing now LayoutTestController::overridePreference method for EFL https://bugs.webkit.org/show_bug.cgi?id=83007. I've noticed that each ewk_view related setting (web preference) is currently present in ewk_view as a separate function (see functions starting with ewk_view_setting_... ) In order to implement LayoutTestController::overridePreference I will need to add more ewk_view_setting_ functions to the ewk_view (as not all the settings are present there). That will affect the existing API and will look quite ugly. Don't you think it would be better to have two common methods (getter and setter) for all the web preferences like : Eina_Bool ewk_view_setting_set(Evas_Object *o, const char* property, const Eina_Value* value ); and Eina_Bool ewk_view_setting_get(const Evas_Object *o, const char* property, Eina_Value* value ); I could do such refactoring within the bug83007. Could you please tell your opinion about my proposal? BR, Mikhail --------------------------------------------------------------------- Intel Finland Oy Registered Address: PL 281, 00181 Helsinki Business Identity Code: 0357606 - 4 Domiciled in Helsinki This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.

9 13

64-bit release buildbot is sick.
by Gyuyoung Kim 23 May '12

23 May '12

Hello webkit-efl folks, As I notify to webkit-dev, I have upgraded efl ews and 64-bit release buildbot. However, unfortunately, there are some problems. Though buildbot succeeds to be upgraded to 12.04, I lost root permission. I will do my best to solve this problem. It looks this problem caused by ubuntu upgrading via ssh. I will let you guys know when I fix this problem. I restore EFL EWS first for now. Gyuyoung

2 2

Structuring test_expectations.txt into categories
by Dominik Röttsches 22 May '12

22 May '12

Hi, I have a proposed patch for organising our test_expectations.txt into several categories, similar to the GTK approach. Your feedback on the related bug https://bugs.webkit.org/show_bug.cgi?id=86628 is welcome. // We divide our test_expectations.txt into several categories: // 1. TESTS EXPECTED TO FAIL // Test cases related to features that we don't expect to be activated for EFL since e.g. // * the feature implementation in WebCore is not complete yet // * the feature is not relevant for the EFL port // * the feature is meant to solve a backwards compatibility issue on other platforms // Test cases that don't apply for the EFL platform permanently. // If in doubt, please discuss on webkit-efl mailing list whether a set of test cases // should be moved here. // 2. CRASHES // 3. FLAKY TESTS // 4. FAILING TESTS // * Interim failures that are moved here for further investigation as a result of daily // gardening. They should be analysed quickly and either resolved or moved to 1). // * Remaining failures that are moved here until further investigation after // unskipping a directory or block of tests in the Skipped list. // * Tests that we expect to pass once the respective fix for DRT is available or // a relevant compile time feature can be enabled for EFL. Regards, Dominik -- Dominik Röttsches<dominik.rottsches(a)intel.com>

2 2

Missing dbus-1 package on debug buildbot
by Gyuyoung Kim 17 May '12

17 May '12

Hello Dominik, EFL Debug build bot is sick because of missing dbus-1 package. Could you please install dbus-1 to debug buildbot? "No package 'dbus-1' found" The dbus-1 package is required by Bug 83254. Cheers, Gyuyoung

2 1

Do we really need a global history implementation?
by Raphael Kubo da Costa 16 May '12

16 May '12

Hey there. Mikhail has been working on a few patches [1][2][3][4] adding support for a global history to the port. Most of the code in those patches looks fine, but before commenting on the bugs I would like to discuss whether we actually need a global history implementation at all. Only the Apple ports (mac and win) really use the global history callbacks in WebCoreSupport (Qt implements some bits just to pass the layout tests [5]). For us, it means we need to add a few more data structures and emit a few very specialized signals I doubt users are going to need (such as "perform,client,redirect" or "navigate,with,data"). The only benefit I see so far is that a few more tests will be unskipped, but then if we choose we do not want to support this feature we can just move the skipped tests to the "Tests which are expected to fail" section. Does anyone have a different opinion or has anyone needed a global history implementation when using the port? [1] https://bugs.webkit.org/show_bug.cgi?id=86343 [2] https://bugs.webkit.org/show_bug.cgi?id=82579 [3] https://bugs.webkit.org/show_bug.cgi?id=86370 [4] https://bugs.webkit.org/show_bug.cgi?id=82724 [5] http://trac.webkit.org/browser/trunk/Source/WebKit/qt/WebCoreSupport/FrameL…

3 3

Buildbot Green
by Dominik Röttsches 10 May '12

10 May '12

Dear WebKit EFL folks, Thanks to a great team effort, the WebKit EFL Linux 64bit Debug Buildbot is now green and http://wkb.ug/85601 - [EFL] Green Tree task force was marked RESOLVED FIXED, see also [1]. A big thanks to everyone involved who contributed patches, investigated and gardened test cases, reviewed and committed our patches. This is a great milestone on the way of improving the stability and robustness of the EFL port. It gives us the powerful opportunity to use run-webkit-tests locally and expect close to zero failures when testing patches during development. That's why we'd like to encourage everybody to make use of this tool and check patches for regressions before landing. If you have any questions regarding how to get the same results on your machine as on the bot, please feel free to post to this list and we'll try to resolve such issues. From now, we'll keep a close eye on the bot and we'll set up a gardening schedule in order to keep the bot green in the future. Dominik [1] http://goo.gl/ei1gL ;-)

1 0

glib integration and timers precision
by Thiago Marcos P. Santos 09 May '12

09 May '12

Hi Webkit EFL, We have the ecore main loop pumping glib main loop because of libsoup (mainly) on EWK. I have the feeling that we are having more issues with timeouts/timers/delays than other platforms and it might be somehow related (noticed that when fixing some flaky tests). - Does anyone knows the performance impact of this pumping? We are also using the default ecore timer precision which has a small delay. - Is it a good idea to set it to maximum precision (maybe on DRT only)? Cheers,

1 0

Ubuntu 12.04 on Buildbots?
by Dominik Röttsches 05 May '12

05 May '12

Hi Gyuyoung, you mentioned you were planning to upgrade the Release buildbot to Ubuntu 11.x Now that Ubuntu 12.04 got released - that doesn't seem like an up-to-date idea anymore. How about skipping 11.x and upgrade it to 12.04 directly? Then we could do the same with the Debug buildbot and they would be in sync. What do you think? Dominik -- Dominik Röttsches<dominik.rottsches(a)linux.intel.com>

4 4