see patch
Created attachment 73833 [details] Patch
*** Bug 49229 has been marked as a duplicate of this bug. ***
Comment on attachment 73833 [details] Patch Thanks for fixing this. The problem must have appeared after http://trac.webkit.org/changeset/71515 where I reasoned that ENABLED(INSPECTOR) guards can be removed from the ConsoleMessage.cpp because the latter is only used by the inspector code itself. I confined my check to compilation of --minimal build which is usually enough to see if there are any issues related to ENABLED(INSPECTOR) macro.
Comment on attachment 73833 [details] Patch Clearing flags on attachment: 73833 Committed r71981: <http://trac.webkit.org/changeset/71981>
All reviewed patches have been landed. Closing bug.
(In reply to comment #3) > (From update of attachment 73833 [details]) > Thanks for fixing this. The problem must have appeared after http://trac.webkit.org/changeset/71515 where I reasoned that ENABLED(INSPECTOR) guards can be removed from the ConsoleMessage.cpp because the latter is only used by the inspector code itself. I confined my check to compilation of --minimal build which is usually enough to see if there are any issues related to ENABLED(INSPECTOR) macro. --minimal option disables features defined in @features in build-webkit script, but INSPECTOR isn't in it. Do we need to it to the build systems? I checked ENABLE_INSPECTOR guards and only PLATFORM(IOS) (in Platform.h) and PLATFORM(ANDROID) (in WebCore/config.h) disable it. To add/remove a feature is so complex task, because all build system have its own default value for all features. And there are platform specific default values in build-webkit script, Platform.h, JavascriptCore/config.h and WebCore/config.h . See http://trac.webkit.org/changeset/71754 for details and https://bugs.webkit.org/show_bug.cgi?id=48755 why is so complex to modify MSVC build files.
*** Bug 49462 has been marked as a duplicate of this bug. ***
Its a little weird having just the implementation of ScriptCallStack::buildInspectorObject being ifdef'd out, but not the signature in the .h. I think the entire file could have been guarded, but this works fine. Thanks for fixing this.
(In reply to comment #8) > Its a little weird having just the implementation of ScriptCallStack::buildInspectorObject being ifdef'd out, but not the signature in the .h. bug 49229#c2 says: > The approach so far has been to flag the dependencies only in the body so that the flag does not bleed over. I'd like to continue this approach. > I think the entire file could have been guarded, but this works fine. You can't ifdef the whole Script*.h files, because there are dependencies in Console.cpp.
Good to know. Thanks!