Netscape Plugin and MediaQueries depend on ecore_x functionality, therefore, ecore_x should be initialized when process starts to avoid re-initialization.
Created attachment 170134 [details] Patch 1 - initialize ecore_x in WebProcess - removed old workaround for ecore_x_init
Comment on attachment 170134 [details] Patch 1 View in context: https://bugs.webkit.org/attachment.cgi?id=170134&action=review > Source/WebKit2/WebProcess/efl/WebProcessMainEfl.cpp:71 > + if (!ecore_x_init(0)) { Where do we shutdown those libs? > Source/WebKit2/WebProcess/efl/WebProcessMainEfl.cpp:75 > + ecore_shutdown(); Missing eina_shutdown()
Comment on attachment 170134 [details] Patch 1 View in context: https://bugs.webkit.org/attachment.cgi?id=170134&action=review > Source/WebKit2/ChangeLog:9 > + PlatformScreenEfl functions and systemBeep() depend on ecore_x functionality, therefore, > + ecore_x should be initialized when web process starts to avoid re-initialization. Does using ecore_x buy os anything than using X directly? and potentially share code with chrome/linux
Created attachment 170137 [details] Patch 2 - Applied fixes according to review comments
Comment on attachment 170137 [details] Patch 2 r=me though I would rather share the Chrome WebScreenInfo code than use ecore_x
(In reply to comment #3) > (From update of attachment 170134 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=170134&action=review > > > Source/WebKit2/ChangeLog:9 > > + PlatformScreenEfl functions and systemBeep() depend on ecore_x functionality, therefore, > > + ecore_x should be initialized when web process starts to avoid re-initialization. > > Does using ecore_x buy os anything than using X directly? and potentially share code with chrome/linux That might be good idea to add shared code for the ports that use X. I think we need to discuss about that on mailing list and ask opinion from chromium folks.
Comment on attachment 170137 [details] Patch 2 Clearing flags on attachment: 170137 Committed r132212: <http://trac.webkit.org/changeset/132212>
All reviewed patches have been landed. Closing bug.