With the patch I'm going to attach to bug 584474, various build errors show up. The ones related to ipc/ are the following: sys_string_conversions_win.obj : error LNK2005: "class std::basic_string<unsigned short,struct std::char_traits<unsigned short>,class std::allocator<unsigned short> > __cdecl base::SysNativeMBToWide(class StringPiece const &)" (?SysNativeMBToWide@base@@YA?AV?$basic_string@GU?$char_traits@G@std@@V?$allocator@G@2@@std@@ABVStringPiece@@@Z) already defined in StringUtil.obj sys_string_conversions_win.obj : error LNK2005: "class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > __cdecl base::SysWideToNativeMB(class std::basic_string<unsigned short,struct std::char_traits<unsigned short>,class std::allocator<unsigned short> > const &)" (?SysWideToNativeMB@base@@YA?AV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@ABV?$basic_string@GU?$char_traits@G@std@@V?$allocator@G@2@@3@@Z) already defined in StringUtil.obj sys_string_conversions_win.obj : error LNK2005: "class std::basic_string<unsigned short,struct std::char_traits<unsigned short>,class std::allocator<unsigned short> > __cdecl base::SysUTF8ToWide(class StringPiece const &)" (?SysUTF8ToWide@base@@YA?AV?$basic_string@GU?$char_traits@G@std@@V?$allocator@G@2@@std@@ABVStringPiece@@@Z) already defined in StringUtil.obj sys_string_conversions_win.obj : error LNK2005: "class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > __cdecl base::SysWideToUTF8(class std::basic_string<unsigned short,struct std::char_traits<unsigned short>,class std::allocator<unsigned short> > const &)" (?SysWideToUTF8@base@@YA?AV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@ABV?$basic_string@GU?$char_traits@G@std@@V?$allocator@G@2@@3@@Z) already defined in StringUtil.obj sha2.obj : error LNK2001: unresolved external symbol "void __cdecl nss::SHA256_Update(struct SHA256ContextStr *,unsigned char const *,unsigned int)" (?SHA256_Update@nss@@YAXPAUSHA256ContextStr@@PBEI@Z) sha2.obj : error LNK2001: unresolved external symbol "void __cdecl nss::SHA256_Begin(struct SHA256ContextStr *)" (?SHA256_Begin@nss@@YAXPAUSHA256ContextStr@@@Z) sha2.obj : error LNK2001: unresolved external symbol "void __cdecl nss::SHA256_End(struct SHA256ContextStr *,unsigned char *,unsigned int *,unsigned int)" (?SHA256_End@nss@@YAXPAUSHA256ContextStr@@PAEPAII@Z) chrome_paths_win.obj : error LNK2001: unresolved external symbol "unsigned short const * const chrome::kBrowserAppName" (?kBrowserAppName@chrome@@3QBGB) chrome_paths_win.obj : error LNK2001: unresolved external symbol "unsigned short const * const chrome::kUserDataDirname" (?kUserDataDirname@chrome@@3QBGB) Since fakelibs doesn't build intermediate libs, the final libxul link might not remove the unnecessary functions that are using these symbols (That may be related to the fact that the final libxul link doesn't do LTO, actually). The first set of errors (duplicate symbols) should be just safe to fix in all cases. The nss symbols are weird, because the sha512.cc file is being compiled for mac and linux, but not windows, and the sha2.cc file is compiled on all platforms.

http://hg.mozilla.org/mozilla-central/rev/848533d01b60 http://hg.mozilla.org/mozilla-central/rev/80d4345705be http://hg.mozilla.org/mozilla-central/rev/e0ac7ffcc68b