I just ran through https://github.com/nodejs/node/blob/master/doc/guides/maintaining-zlib.md#updating-zlib, and it didn't produce a buildable zlib, see below.
This isn't an immediate problem, but if we have to update zlib for any reason (sec release, performance, portability), whoever is doing that will find the docs not quite sufficient.
If anyone wants to figure out what it would take to fix the docs, it'd be good to do it sooner than later.
Also, it might be useful to extend the docs to see how to determine what upstream SHA has actually been released in a chromium, so we don't grab pre-release/untested code.
cc: @mhdawson @richardlau @MylesBorins (to whom I mentioned that I don't think the guide is still in working order).
[19/1004] CC obj/deps/zlib/zlib.adler32_simd.o
FAILED: obj/deps/zlib/zlib.adler32_simd.o
clang-11 -MMD -MF obj/deps/zlib/zlib.adler32_simd.o.d -DV8_DEPRECATION_WARNINGS -DV8_IMMINENT_DEPRECATION_WARNINGS -D__STDC_FORMAT_MACROS -DOPENSSL_NO_PINSHARED -DOPENSSL_THREADS -DHAVE_HIDDE
N -DADLER32_SIMD_SSSE3 -DINFLATE_CHUNK_SIMD_SSE2 -DCRC32_SIMD_SSE42_PCLMUL -DINFLATE_CHUNK_READ_64LE -I../../deps/zlib -pthread -Wall -Wextra -Wno-unused-parameter -m64 -O3 -fno-omit-frame-po
inter -c ../../deps/zlib/adler32_simd.c -o obj/deps/zlib/zlib.adler32_simd.o
../../deps/zlib/adler32_simd.c:116:34: error: always_inline function '_mm_maddubs_epi16' requires target feature 'ssse3', but would be inlined into function 'adler32_simd_' that is compiled w
ithout support for 'ssse3'
const __m128i mad1 = _mm_maddubs_epi16(bytes1, tap1);
^
../../deps/zlib/adler32_simd.c:120:34: error: always_inline function '_mm_maddubs_epi16' requires target feature 'ssse3', but would be inlined into function 'adler32_simd_' that is compiled w
ithout support for 'ssse3'
const __m128i mad2 = _mm_maddubs_epi16(bytes2, tap2);
^
2 errors generated.
[24/1003] CC obj/tools/icu/gen/icudata.icudt66_dat.o
I just ran through https://github.com/nodejs/node/blob/master/doc/guides/maintaining-zlib.md#updating-zlib, and it didn't produce a buildable zlib, see below.
This isn't an immediate problem, but if we have to update zlib for any reason (sec release, performance, portability), whoever is doing that will find the docs not quite sufficient.
If anyone wants to figure out what it would take to fix the docs, it'd be good to do it sooner than later.
Also, it might be useful to extend the docs to see how to determine what upstream SHA has actually been released in a chromium, so we don't grab pre-release/untested code.
cc: @mhdawson @richardlau @MylesBorins (to whom I mentioned that I don't think the guide is still in working order).