A hub header is one that is included by many other headers, and itself includes many other headers. See tracking bug 912735.

Er, this instead: https://tbpl.mozilla.org/?tree=Try&rev=7d07cbba51a7

Windows-only: https://tbpl.mozilla.org/?tree=Try&rev=8874400b7477

http://hg.mozilla.org/integration/mozilla-inbound/rev/d39afeecef66 http://hg.mozilla.org/integration/mozilla-inbound/rev/eb9ce350a4ad

Looks like SaturatingInflate isn't used anymore, so it should probably be removed: http://mxr.mozilla.org/mozilla-central/search?string=SaturatingInflate SaturatingUnionEdges is used heavily on performance critical paths in layout though, and I thought we intentionally made it inline for that reason.

https://hg.mozilla.org/mozilla-central/rev/d39afeecef66 https://hg.mozilla.org/mozilla-central/rev/eb9ce350a4ad

(In reply to Mats Palmgren (:mats) from comment #8) > Looks like SaturatingInflate isn't used anymore, so it should probably be > removed: > http://mxr.mozilla.org/mozilla-central/search?string=SaturatingInflate Good point! > > SaturatingUnionEdges is used heavily on performance critical paths in layout > though, > and I thought we intentionally made it inline for that reason. According to DXR, it's not used anymore either: http://dxr.mozilla.org/mozilla-central/search?q=SaturatingUnionEdges&redirect=false

Ah, no, I see, this is used by UnionEdges, which is indeed used outside of this file: http://dxr.mozilla.org/mozilla-central/search?tree=mozilla-central&q=UnionEdges&redirect=true

Comment on attachment 805129 [details] [diff] [review] Address Mats' comments > +#include <algorithm> // for algorithm s/for algorithm/for min, max/ and we should move back this include from .cpp to .h: #include "nsDebug.h" // for NS_WARNING and remove #include <algorithm> in .cpp

Thanks, integrated your comments, new try push https://tbpl.mozilla.org/?tree=Try&rev=242e7bf50719

Comment on attachment 805129 [details] [diff] [review] Address Mats' comments [Approval Request Comment] Bug caused by (feature/regressing bug #): 913603 User impact if declined: possible performance regression Testing completed (on m-c, etc.): m-i Risk to taking this patch (and alternatives if risky): none (reverting a recent change) String or IDL/UUID changes made by this patch: none