v7.0.0 Proposal

[jasnell]

• Buffer
  • Passing invalid input to Buffer.byteLength will now throw an error #8946.
  • Calling Buffer without new is now deprecated and will emit a process warning #8169.
  • Passing a negative number to allocUnsafe will now throw an error #7079.
• Child Process
  • The fork and execFile methods now have stronger argument validation #7399.
• Cluster
  • The worker.suicide method is deprecated and will emit a process warning #3747.
• Deps
  • V8 has been updated to 5.4.500.31 #8852.
  • NODE_MODULE_VERSION has been updated to 51 #8808.
• File System
  • A process warning is emitted if a callback is not passed to async file system methods #7897.
• Intl
  • Intl.v8BreakIterator constructor has been deprecated and will emit a process warning #8908.
• Promises
  • Unhandled Promise rejections have been deprecated and will emit a process warning #8217.
• Punycode
  • The punycode module has been deprecated #7941.
• URL
  • An Experimental WHATWG URL Parser has been introduced #7448.

[targos]

Copying over my comment from #8503 (comment):

I have a draft of a document explaining the new features and performance improvements that are coming with V8 5.4: https://gist.github.com/targos/eda0bd7068e38b5ece5802aaefa6e601

I don't know if this should be in the changelog, on a blog post or something else ?

[jasnell]

Likely will include a short statement in the changelog but will provide more detail in a blog post

[dan-mcdonald]

Looks like this commit ID is missing back ticks, i.e. should be 8a87b29034

[jasnell]

Beta build run: https://ci-release.nodejs.org/job/iojs+release/1272/
This will be the final beta. The plan is to cut the first release candidate on Thursday, followed by a second release candidate as necessary on Monday, with the release targeted for Tuesday, October 25

[jasnell]

Beta build artifacts are here: https://nodejs.org/download/test/v7.0.0-test201610180ccb98695e/
This build includes the release bit flipped as this is the final build before the actual release. This is only intended for a final round of testing. The first release candidate should arrive on Thursday.

/cc @nodejs/testing @nodejs/build

[jbergstroem]

I'm going to proceed with nodejs/build#500 ahead of Thursday unless anyone has any objections.

[rvagg]

@jasnell would you mind rebasing and editing one of my commits that didn't get metadata on master:

    tools: explicitly set digest algo for SHASUM to 256

should have

    PR-URL: https://github.com/nodejs/node/pull/9071
    Reviewed-By: James M Snell <jasnell@gmail.com>
    Reviewed-By: Myles Borins <myles.borins@gmail.com>

I've done this on v4.x-staging

[rvagg]

@jasnell I've cherry-picked these into v7.x-staging, they were the relevant security updates that went into v6.9.0 that belong here too:

• cae9eb3 crypto: fix openssl.cnf FIPS handling & testing
• 1ea0358 node: --openssl-config cli argument
• 647afe9 inspector: generate UUID for debug targets

[jasnell]

@nodejs/ctc @nodejs/collaborators ... v7.0.0 is almost ready to go. Please take a moment to look over the changelog and let me know if anything looks amiss or should be called out in the notable changes.

[jasnell]

@rvagg ... I fixed that commit and pulled in the security fixes.

[jasnell]

CITGM: https://ci.nodejs.org/view/Node.js-citgm/job/citgm-smoker/426/
CI: https://ci.nodejs.org/job/node-test-pull-request/4570/

[silverwind]

Probably not relevant to the release, but those hash links in your initial post don't work.

[benjamingr]

I'd like to change the wording on unhandled promise rejections.

A few members of the promises community have concerns about that wording and I tend to agree.

Can we go with "promise rejections not handled within the same tick will now emit a warning on the process object unless an "unhandledRejection" event handler is installed"?

[targos]

I'd also add a reference to #8317 in the V8 update line. 5.4.500.31 was minor.

[rvagg]

@benjamingr if you think the wording on the post(s) at nodejs.org and/or in the changelogs could be improved around that then just file a pull request to adjust it.

[jbergstroem]

sunos builds should now show up as smartos.

[jasnell]

@nodejs/collaborators ... all, I'm going to be kicking off the first release candidate build in a couple of minutes. This means that after this point, unless the CTC deems it absolutely necessary, I won't be landing any more semver-minor or semver-major commits into v7.0.0. The second RC build will go out on Monday.

[jasnell]

RC.1 Build: https://ci-release.nodejs.org/job/iojs+release/1283/ (cancelled)
RC.1 Build: https://ci-release.nodejs.org/job/iojs+release/1285/ (argh! failed because I didn't wait long enough before kicking it off.... will try again shortly)
RC.1 Build (third time's the charm): https://ci-release.nodejs.org/job/iojs+release/1286/
Nope, that one hiccuped also... number 4: https://ci-release.nodejs.org/job/iojs+release/1287/

Looks like there's some issues with the release job... waiting until those are resolved before continuing...

alright... finally got a good one... almost done: https://ci-release.nodejs.org/job/iojs+release/1289/

@nodejs/testing RC.1 artifacts are here: https://nodejs.org/download/rc/v7.0.0-rc.1/

[jasnell]

Will be landing #9253 soon and will be wrapping this up and putting a bow on it. Will do the actual release build tonight and will promote in the morning.

[seishun]

Could we still get the libuv upgrade in? /cc @cjihrig

[jasnell]

It's a bit too late for that, I'm afraid, There's no time to adequately test it. My plan is to do a quick spin around on a v7.1 next week to get the libuv and icu updates in

[jasnell]

RC.2 Build: https://ci-release.nodejs.org/job/iojs+release/1294/

[jasnell]

RC.2 test artifacts are here: https://nodejs.org/download/rc/v7.0.0-rc.2/

[jbergstroem]

just a fyi: ran into issues renaming smartos for rc1; will do another attempt soon.

[jasnell]

Release build is underway

[jasnell]

merged. just waiting on the website to refresh now!