Comment on attachment 813136 [details] [diff] [review] Part 2: Don't #include nsIDocument.h in mozilla/dom/Element.h Review of attachment 813136 [details] [diff] [review]: ----------------------------------------------------------------- I'd rather check if we can move them out of line altogether... The JS guys just figured out what a footgun their -inl.h files are, and started removing them. ::: content/base/public/ElementInlines.h @@ +60,5 @@ > + > +} > +} > + > +#endif // ElementInlines_h Comment is wrong

(In reply to comment #2) > I'd rather check if we can move them out of line altogether... The JS guys just > figured out what a footgun their -inl.h files are, and started removing them. Can you be more specific please? (Also note that this won't be the first "-inl.h" file we have in Gecko, so addressing those issues may be beyond the purpose of this bug.) Also, in my experience, the compiler basically ignores whether a function is declared as inline or not, and makes its own inlining decisions. In presence of PGO, it pretty much does what it wants to the code. :-)

https://hg.mozilla.org/integration/mozilla-inbound/rev/ec20bcd7c902 https://hg.mozilla.org/integration/mozilla-inbound/rev/bfaaa5b35899

https://hg.mozilla.org/mozilla-central/rev/ec20bcd7c902 https://hg.mozilla.org/mozilla-central/rev/bfaaa5b35899