I just hit this locally on windows with clang/win:




[1044/32652] CC obj\third_party\boringssl\src\crypto\boringssl.time_support.obj
FAILED: ninja -t msvc -e environment.x86 -- "..\..\third_party/llvm-build/Release+Asserts/bin/clang-cl" -m32 /nologo /sh
owIncludes /FC @obj\third_party\boringssl\src\crypto\boringssl.time_support.obj.rsp /c ..\..\third_party\boringssl\src\c
rypto\time_support.c /Foobj\third_party\boringssl\src\crypto\boringssl.time_support.obj /Fdobj\third_party\boringssl\bor
ingssl.c.pdb
In file included from ..\..\third_party\boringssl\src\crypto\time_support.c:62:
..\..\third_party\boringssl\src\include\openssl/time_support.h(73,33) :  error: unknown type name 'time_t'; did you mean
 'size_t'?
struct tm *OPENSSL_gmtime(const time_t *timer, struct tm *result);


I tweaked my local header setup a bit for working on  bug 592745 , so this is only broken on my machine -- but it's a true positive: time_support.h doesn't include anything but uses time_t.  Either it should include time.h, or all of its clients need to include time.h before including time_support.h. time_support.c doesn't.

davidben: If you tell me which of the two patches to write, I can write the patch if you want.