Good news, they have finally fixed the EVP message!

Another good news: keypress patch should be no longer floated, because it is included in 1.0.1k!

@bnoordhuis: please check that your x32 changes are still in. I have used opensslconf from the origin/v1.x, so I assume - yes.

Can you make the CI check this PR? I suspect it's going to break the Windows build because of the removal of (at least) x86_64-win32-masm.asm.

@bnoordhuis : https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/48/

@bnoordhuis: appears to be building fine on windows.

@indutny gyp is definitely complaining:

Warning: Missing input files: deps\openssl\openssl\crypto\bn\asm\x86_64-win32-masm.asm

The fact that it builds is presumably because the Windows bots build 32 bits binaries only. Maybe @rvagg can confirm.

This PR also seems to break parallel/test-https-foafssl on Windows.

@bnoordhuis: pushed update for GYP.

@bnoordhuis : regarding foafssl - it is not completely evident to me that this is caused by this PR.

@indutny Would you mind also updating https://github.com/iojs/io.js/blob/v1.x/deps/openssl/asm/Makefile? I noticed some asm files er axed in this release.

Building x64 on windows:

openssl.lib(bn_word.obj) : error LNK2001: unresolved external symbol bn_mul_wor
ds [D:\iojs\deps\openssl\openssl-cli.vcxproj]
openssl.lib(bn_word.obj) : error LNK2001: unresolved external symbol bn_div_wor
ds [D:\iojs\deps\openssl\openssl-cli.vcxproj]
openssl.lib(bn_add.obj) : error LNK2001: unresolved external symbol bn_add_word
s [D:\iojs\deps\openssl\openssl-cli.vcxproj]
openssl.lib(bn_mont.obj) : error LNK2001: unresolved external symbol bn_mul_add
_words [D:\iojs\deps\openssl\openssl-cli.vcxproj]
openssl.lib(bn_mont.obj) : error LNK2001: unresolved external symbol bn_sub_wor
ds [D:\iojs\deps\openssl\openssl-cli.vcxproj]
openssl.lib(bn_mul.obj) : error LNK2001: unresolved external symbol bn_mul_comb
a8 [D:\iojs\deps\openssl\openssl-cli.vcxproj]
openssl.lib(bn_mul.obj) : error LNK2001: unresolved external symbol bn_mul_comb
a4 [D:\iojs\deps\openssl\openssl-cli.vcxproj]
openssl.lib(bn_sqr.obj) : error LNK2001: unresolved external symbol bn_sqr_comb
a4 [D:\iojs\deps\openssl\openssl-cli.vcxproj]
openssl.lib(bn_sqr.obj) : error LNK2001: unresolved external symbol bn_sqr_comb
a8 [D:\iojs\deps\openssl\openssl-cli.vcxproj]
openssl.lib(bn_sqr.obj) : error LNK2001: unresolved external symbol bn_sqr_word
s [D:\iojs\deps\openssl\openssl-cli.vcxproj]
D:\iojs\Release\openssl-cli.exe : fatal error LNK1120: 10 unresolved externals
[D:\iojs\deps\openssl\openssl-cli.vcxproj]

@piscisaureus could you please try it again with a new commit?

Trying

@indutny yes, that works

WRT that missing file - I wrote that myself one day when I was really bored: 3568edf. I don't mind axing it; since openssl now seems properly maintained I no longer feel the need to outsmart them.

@piscisaureus do you want me to revert it?

@indutny no, it works. That asm file I created was derived from the gcc asm implementation, but since that file has been updated since it seems safer to exclude my own version. Go ahead and land.

Ok, any objections @bnoordhuis ?

Well, I still don't get why parallel/test-https-foafssl is failing now when it was passing before.

@bnoordhuis let's give it another spin on jenkins? https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/49/

@bnoordhuis seems to be timeout-ing anyway. @piscisaureus does it terminate when you run it outside of the test suite? (./iojs test/parallel/test-https-foafssl.js)

@indutny

That'd be Release\iojs.exe test/parallel/test-https-foafssl.js and yes, it hangs.

brave enough to land this prior to 1.0.0 or should we hold off?

@piscisaureus You mean it hangs with or without this PR applied? If it's without, then the failure is not a regression and this PR should be good to go.

Going to see it myself.

Stupid me, I thought that the patch was included, but it was the artifact of cherry-pick which has disappeared after cherry-pick --abort. Here is the latest build: https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/53/

jenkins is good with it now. @bnoordhuis could you please land it if it LGTY?

I carefully scrutinized each and every one of those +1,984 −5,363 changed lines. LGTM!

Landed in eebdf7a, b910613, ced41b0, a76811c, 7c4a50d ! Thanks everyone!