QtWidgets dependencies must be eliminated to compile wtf/jsc/webcore without them.
The goal of this bug is now to make it possible to build WebKit without QtWidgets, with the resulting library not containing the Qt WK1 APIs that depend on QtWidgets.
Next up is moving the QStyle based themeing code out of WebCore, so that we can compile WebCore always without widgets (to avoid accidentally introducing new dependencies there).
Is this fixed now? All dependencies are closed :)
(In reply to comment #3) > Is this fixed now? All dependencies are closed :) What's left is making it possible to build the entire library without QtWidgets, which would effectively disable the WebKit1 API and just leave us with a library that implements the WK2 QML element that is made accessible to applications via the QML plugin. The change is I think largely a build system change to not compile WK1 files if there are no widgets.
CC'ing Tor Arne for upcoming build system surgery :)
Created attachment 149323 [details] Patch
Comment on attachment 149323 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=149323&action=review r=me with a few minor nitpicks > Source/api.pri:141 > QMAKE_INTERNAL_INCLUDED_FILES *= WebKit2/Target.pri You need the same hack for WebKit1.pri > Tools/qmake/mkspecs/features/default_pre.prf:108 > + root_project_file: error("Building WebKit against Qt 5.0 requires the QtWidgets and QtPrintSupport modules. Disabling WebKit1.") This is no longer a fatal error and should use warning instead of error. The message could also reflect this, eg, "WebKit1 requires the QtWidgets and QtPrintSupport modules. Disabling WebKit1."
(In reply to comment #7) > (From update of attachment 149323 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=149323&action=review > > r=me with a few minor nitpicks Thanks for the quick review :) > > Source/api.pri:141 > > QMAKE_INTERNAL_INCLUDED_FILES *= WebKit2/Target.pri > > You need the same hack for WebKit1.pri Good catch! > > Tools/qmake/mkspecs/features/default_pre.prf:108 > > + root_project_file: error("Building WebKit against Qt 5.0 requires the QtWidgets and QtPrintSupport modules. Disabling WebKit1.") > > This is no longer a fatal error and should use warning instead of error. The message could also reflect this, eg, "WebKit1 requires the QtWidgets and QtPrintSupport modules. Disabling WebKit1." Oops, excellent point. I had that commented out while testing ;)
Committed r121228: <http://trac.webkit.org/changeset/121228>
(In reply to comment #9) > Committed r121228: <http://trac.webkit.org/changeset/121228> FYI: It broke the Windows build. r121227 builds fine, and r121230 is broken. You can find a full clean build log for r121230 here: http://build.webkit.org/builders/Qt%20Windows%2032-bit%20Debug/builds/45144
(In reply to comment #10) > (In reply to comment #9) > > Committed r121228: <http://trac.webkit.org/changeset/121228> > > FYI: It broke the Windows build. r121227 builds fine, and r121230 is broken. > You can find a full clean build log for r121230 here: > http://build.webkit.org/builders/Qt%20Windows%2032-bit%20Debug/builds/45144 Yes. I'm working on a fix in bug #88873. I expected exactly this to happen :)
*** Bug 88168 has been marked as a duplicate of this bug. ***