+1 !!! Thanks for quick release.

+1 just for the goto fail additions

how safe is this to land straight into 1.0.0 @indutny? I don't have a feel for how major (and potentially dangerous) these changes are.

@rvagg I feel safe about it, every commit was reviewed by either me or @bnoordhuis

ok LGTM then

Landed in d790f61, thank you.

I just benchmarked using benchmark/http.sh .
http is improved! Thanks.

machine spec:
CPU : 2.3 GHz Intel Core i7
Memory : 16GB
OSX Yosemite

Sorry, this is going to be postponed until next release. It breaks build on windows and some tests became flacky (according to users).

hm, no problem. i am looking forward to next release.

Test fail with http_parser 2.4.0 on OS X:

=== release test-http-1.0 ===                                                  
Path: parallel/test-http-1.0
assert.js:100
  throw new assert.AssertionError({
        ^
AssertionError: 1 == 2
    at response_validator (/private/tmp/io.js/test/parallel/test-http-1.0.js:154:12)
    at Socket.<anonymous> (/private/tmp/io.js/test/parallel/test-http-1.0.js:48:7)
    at Socket.emit (events.js:117:20)
    at _stream_readable.js:888:16
    at process._tickCallback (node.js:337:11)
Command: out/Release/iojs /private/tmp/io.js/test/parallel/test-http-1.0.js

@jbergstroem is it always failing? Appears to be working on my side.

@indutny Just tried once (as well as after pulling the revert). Let me check.

Edit: Nope, can't reliably reproduce.

@jbergstroem I think this is just a test problem. I built http-parser with ASAN, and sanitizers for various overflows, and got no errors and no test failures.

@indutny I haven't seen it since; don't take it into consideration until I can give reliable info.