gfx/qcms/qcmstypes.h:39:23: error: typedef redefinition with different type s ('unsigned long' vs '__uintptr_t' (aka 'unsigned int')) typedef unsigned long uintptr_t; ^ /usr/include/sys/_stdint.h:78:22: note: previous definition is here typedef __uintptr_t uintptr_t; ^ /usr/include/sys/_stdint.h:78:22: note: previous definition is here typedef __uintptr_t uintptr_t; ^ 1 error generated. Technically a regression from bug 594769 but every system tends to have different guard against the type redefinition. - FreeBSD has _UINTPTR_T_DECLARED since 9.0 while 7.4/8.3 only have _INTPTR_T_DECLARED - NetBSD has uintptr_t - OpenBSD and Solaris guard whole header only http://gitweb.dragonflybsd.org/dragonfly.git/blob/HEAD:/include/stdint.h http://svnweb.freebsd.org/base/head/sys/sys/_stdint.h http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/sys/stdint.h http://www.openbsd.org/cgi-bin/cvsweb/src/sys/sys/stdint.h https://github.com/illumos/illumos-gate/blob/master/usr/src/uts/common/sys/int_types.h (included by stdint.h) As a result every BSD patches differently. - FreeBSD fixed in ports http://svnweb.freebsd.org/ports/head/www/firefox/files/patch-gfx-qcms-qcmstypes.h - DragonFly and NetBSD fixed in pkgsrc http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/devel/xulrunner/patches/patch-at - OpenBSD fixed by bug 651444 - Solaris fixed by bug 487765 However, since bug 579517 both added StandardInteger.h (wrapper for stdint.h) neither typedefs for int*_t nor already included headers are needed and the whole inclusion block can be simplified...

Can someone do a Try run?

Comment on attachment 699405 [details] [diff] [review] don't redefine types >-#ifdef __OS2__ >-/* OS/2's stdlib typdefs uintptr_t. So we'll just include that so we don't collide */ >+#elif __OS2__ Why not #elif defined(__OS2__) ?

Hopefully, no more typos.

Try push at https://tbpl.mozilla.org/?tree=Try&rev=2fb1bbf08dfd

I tested it too on OpenBSD without fallout, and all was green on try too. https://hg.mozilla.org/integration/mozilla-inbound/rev/782e3ab94db7