Hi, Have you ever struggled with one of GCC's "used but never defined" warnings? I certainly have, and I've just now worked out a good way to work out what the problem is. Take this example:
../../gc/Barrier.h:489:17: warning: inline function ‘void js::HeapSlot::init(JSRuntime*, JSObject*, js::HeapSlot::Kind, uint32_t, const JS::Value&)’ used but never defined [enabled by default] What I do is rename that function to something that will cause a compile error at its use sites, e.g. init2(). I then recompile. The first compile error is very likely to be the problem. In this case I got the following: ../vm/ObjectImpl-inl.h: In member function ‘void js::ObjectImpl::initializeSlotRange(uint32_t, uint32_t)’: ../vm/ObjectImpl-inl.h:226:85: error: no matching function for call to ‘js::HeapSlot::init(JSRuntime*&, JSObject*, js::HeapSlot::Kind, uint32_t, JS::Value)’ And sure enough, vm/ObjectImpl-inl.h is using HeapSlot::init() but it doesn't #include gc/Barrier-inl.h, which defines HeapSlot::init(). It might not be the first compile error (though in my experience it usually is), but it will be an error in a .h file. Now this doesn't tell you how to fix it, but at least it tells you where the problem is, which is often the hard part (in my experience). Nick ps: jorendorff and I are working on detangling SpiderMonkey's horrific #include hairball. See https://bugzilla.mozilla.org/show_bug.cgi?id=880088 and the dependent bugs if you want to follow along. This should simplify things, and hopefully also reduce build times. _______________________________________________ dev-tech-js-engine-internals mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals
