Release proposal: v1.8.2

[Fishrock123]

Suggesting a fixes release, just landed the first bunch of backports (#1560), going to compiling a PR of more today / tomorrow. Proposed release date: soon, maybe next monday?

• [5404cbc745] - buffer: fix copy() segfault with zero arguments (Trevor Norris) #1520
• [65dd10e9c0] - build: remove -J from test-ci (Rod Vagg) #1544
• [74060bb60e] - crypto: track external memory for SSL structures (Fedor Indutny) #1529
• [f10f379240] - deps: make node-gyp work with io.js (cjihrig) #990
• [ba0e744c2c] - deps: upgrade npm to 2.9.0 (Forrest L Norvell) #1583
• [b3a7da1091] - deps: update http_parser to 2.5.0 (Fedor Indutny) #1517
• [4030545af6] - fs: validate fd on fs.write (Julian Duque) #1553
• [898d423820] - string_decoder: don't cache Buffer.isEncoding (Brian White) #1548
• [32a6dbcf23] - test: extend timeouts for ARMv6 (Rod Vagg) #1554
• [5896fe5cd3] - test: adjust Makefile/test-ci, add to vcbuild.bat (Rod Vagg) #1530
• [b72e4bc596] - tls: destroy singleUse context immediately (Fedor Indutny) #1529
• [1cfc455dc5] - tls: zero SSL_CTX freelist for a singleUse socket (Fedor Indutny) #1529
• [7ada680519] - tls: destroy SSL once it is out of use (Fedor Indutny) #1529
• [71274b0263] - tls_wrap: use localhost if options.host is empty (Guilherme Souza) #1493
• [0eb74a8b6c] - win,node-gyp: optionally allow node.exe/iojs.exe to be renamed (Bert Belder) #1266

@rvagg looks like changelog-maker is still having issues with --start-ref. Using --start-ref=v1.8.1 while on branch v1.x still produces these two extra changes that are out-of-order and already in v1.8.1:

• [53ed89d927] - _Revert_ "build: use %PYTHON% instead of python" (Rod Vagg) #1475
• [f23b96352b] - src: revert NODE_MODULE_VERSION to 43 (Chris Dickinson) #1460

[rvagg]

I'm going to punt on engaging with this thread until at least the weekend, I have too much on my plate till then but I'd like to be involved in this because of the trickiness of the release mechanics and this being the first out-of-order release we've done (i.e. there's some parts of the release process that I know will have problems with this and I'd like to use this opportunity to sort those out if possible).

[Fishrock123]

Ah, ok. There isn't a big rush on this imo, but it should probably happen soon.

[rvagg]

good work!

(edit: derp)

[Fishrock123]

This is for 1.8.2 :)

[rvagg]

I might try and push this through while in (lengthy) transit tomorrow (Sunday USA)

[rvagg]

https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/684/

[Fishrock123]

Ah shoot, I forgot to make a new PR with more backports. I wouldn't do this yet, there's a bunch of other things we should be backporting into it. (I might be able to get that tomorrow... if not, tuesday.) (also need to decide if we want to backport perf enhancements)

[rvagg]

that'll be for v1.8.3 then, worth experimenting with a small delta I think

[rvagg]

https://jenkins-iojs.nodesource.com/job/iojs+release/65/

[Fishrock123]

Ok fair enough.

[rvagg]

sooooo ... that worked perfectly, it seems that I had enough magical foresight when creating all of the release scripting for this to just .. work.

https://iojs.org/dist/v1.8.2/

https://iojs.org/dist/latest/ still points to v2.0.2

index.tab and index.json appear to be correct and have the right order

perhaps @nodejs/website want to consider how they might frame this as an LTS-ish thing, 1.8.2 will continue to get maintenance for a while although we can't give a timeframe until the LTS group comes up with a policy of sorts.