Do we want extra metadata on the cherry-picked commits or not? I may look into how much work making some tooling for it would be..

For the interested, made via this command, since there is nothing semver-major past 4.0.0 bump on master:

git cherry-pick 936c9ffb0fa8e023b40e20882c16aaaef0ba4178^..226b8e0cf8374a0074142145858fa0aa94bca3ce

Do we want extra metadata on the cherry-picked commits or not?

I don't see a case for this, unless you have something in mind?

Also, let's not assume that everything that's not semver-major goes into a release, the list is fine for now but we're going to need to deal with awkward subtleties in the future.

847459c was landed as a major to master?

@brendanashworth that commit isn't listed here. :)

I'd like to see #2352 land (and any other things that might be the result of #2308) in this release.

Also +1 for something like #2352. @Fishrock123 does that mean you're cutting the release before that commit?

@silverwind No, this list was the applicable commits at the time this PR was created, I'll update it as we get closer.

+1 for holding off till the memory leaks are fixed, I'd consider 3.0.0 somewhat broken until we have that sorted out

Could we add these two?

• build: bump ICU from 54 to 55 #2292 6dc01eb (Intl: icu4c bump, no new features - landed in master, ready to go)
• intl: VS2015: workaround ICU VS2015 failure? #2279 VS2015 edit: Withdrawing this one.

@srl295 I'm willing to hold it for the memory leak fixes, but the rest will have to just come as they land; there will probably be weekly-ish releases from stable lines still.

@Fishrock123 No reason to hold. #2292 is ready to merge.

@Fishrock123 Can you include 2781333 - converged Intl? #238

Added:

git cherry-pick ad7f74453dfe4b13738e13eaf039e73c3daf1fd8
git cherry-pick a642a76ef4a5f49eeeaf0668b7a851d8d87e812b^..27813339cf713275fa7bbf2ae497d13b66aa78ce

@Fishrock123 The memory leak PR is up, reviewed and green (#2352). I usually leave contributors to merge their own PR's, but if it's a hurry then it's probably fine to land now.

#2352 just landed in 2781333...8841947.

I'll try to write a changelog soon, but #2375 also looks promising for plugging mem leaks.

One problem with not cherry-picking regularly and letting them bunch up until a release is that we don't get nightlies for the stable branches except when we also get a release. This is something we're going to have to figure out a process for.

In the meantime, because I want a nightly that I can do some memory testing from, I've cherry-picked the latest, excluding the version bump and the semver-major. The commit log looks like this now:

• [3645dc62ed] - build: work around VS2015 issue in ICU <56 (Steven R. Loomis) nodejs/node#2283
• [455ec570d1] - src: rename Buffer::Use() to Buffer::New() (Ben Noordhuis) nodejs/node#2352
• [fd63e1ce2b] - src: introduce internal Buffer::Copy() function (Ben Noordhuis) nodejs/node#2352
• [5586ceca13] - src: move internal functions out of node_buffer.h (Ben Noordhuis) nodejs/node#2352
• [bff9bcddb6] - src: plug memory leaks (Ben Noordhuis) nodejs/node#2352
• [1f12e03266] - (SEMVER-MINOR) build: intl: converge from joyent/node (Steven R. Loomis) nodejs/node#2264
• [071640abdd] - build: Intl: bump ICU4C from 54 to 55 (Steven R. Loomis) nodejs/node#2293
• [e51f90d747] - test: option to run a subset of tests (João Reis) #2260
• [ccf12df4f3] - (SEMVER-MINOR) src: add total_available_size to v8 statistics (Roman Klauke) #2348
• [cc46d3bca3] - test: clarify dropMembership() call (Rich Trott) #2062
• [7118b8a882] - path: remove dead code in favor of unit tests (Nathan Woltman) #2282
• [07a88b0c8b] - build: update manifest to include Windows 10 (Lucien Greathouse) #2332
• [7606bdb897] - (SEMVER-MINOR) util: display constructor when inspecting objects (Christopher Monsanto) #1935
• [94b765f409] - tls: fix check for reused session (Fedor Indutny) #2312
• [e83a41ad65] - tls: introduce internal onticketkeycallback (Fedor Indutny) #2312
• [0ee4df9c7a] - test: make listen-fd-cluster/server more robust (Sam Roberts) #1944
• [1738c9680b] - net: ensure Socket reported address is current (Ryan Graham) #2095
• [cf9ba81398] - test: address timing issues in simple http tests (Gireesh Punathil) #2294
• [e76822f454] - doc: multiple documentation updates cherry picked from v0.12 (James M Snell) #2302
• [0bb099f444] - build: expand ~ in install prefix early (Ben Noordhuis) #2307

@Fishrock123: FYI changelog-maker is getting an upgrade to handle the GH API redirects for nodejs/io.js -> nodejs/node (yay for API redirects!) but the patch process is snaking its way through various modules that I need collaborator signoff on (jsonist -> ghutils -> ghissues -> changelog-maker), should be ready soon but the above is the non-grouped result of running it on v3.x.

Here's the nightly: https://iojs.org/download/nightly/v3.0.1-nightly201508173645dc62ed/

I'll report back if I find anything re memory but don't let that hold up a release since it appears that we have the major issues sorted out. At a glance it doesn't look like #2375 is dealing with anything v3 specific and they mostly look like one-off minor non-frees anyway.

^

git cherry-pick 64577463dfbc7ebe646c6c5a06e4b309b2757205^..4c06515a2f13c9b0890b374bb3ab3c0740c282e2

@rvagg Why do we have the (SEMVER-MINOR) label? With the new way the branches are managed I'm missing the utility in it.

@trevnorris making it easier for us to know when to bump minors in the stable branch. :)

@Fishrock123 yeah... very obvious now. thanks.

@trevnorris (SEMVER-MINOR) is printed on the CHANGELOG list, it comes from the semver-minor GitHub label, no need to change anything in actual commit messages.

Sorry it's been longer, will try to get this out tomorrow morning.

CI: https://jenkins-iojs.nodesource.com/job/node-test-commit/186/

Edit: fixed.

changelog-maker is now updated and follows GitHub API redirects to get the PR labels properly, update your copy @Fishrock123

git cherry-pick a593cb7682df7e45550d1ebe26be310dc7bef6bd^..90a267188937c9f8d522fa6095a1ff8087fa8ee0

New CI: https://jenkins-iojs.nodesource.com/job/node-test-commit/191/

git cherry-pick d98eed51f782ea44c3fd7823b2912f7fb30ab185

• [3645dc62ed] - build: work around VS2015 issue in ICU <56 (Steven R. Loomis) #2283
• [1f12e03266] - (SEMVER-MINOR) build: intl: converge from joyent/node (Steven R. Loomis) #2264
• [071640abdd] - build: Intl: bump ICU4C from 54 to 55 (Steven R. Loomis) #2293
• [07a88b0c8b] - build: update manifest to include Windows 10 (Lucien Greathouse) #2332
• [0bb099f444] - build: expand ~ in install prefix early (Ben Noordhuis) #2307
• [7fe6dd8f5d] - crypto: check for OpenSSL errors when signing (P.S.V.R) #2342
• [605f6ee904] - crypto: fix memory leak in PBKDF2Request (Karl Skomski) #2375
• [ba6eb8af12] - crypto: fix memory leak in ECDH::SetPrivateKey (Karl Skomski) #2375
• [6a16368611] - crypto: fix memory leak in PublicKeyCipher::Cipher (Karl Skomski) #2375
• [a760a87803] - crypto: fix memory leak in SafeX509ExtPrint (Karl Skomski) #2375
• [f45487cd6e] - crypto: fix memory leak in SetDHParam (Karl Skomski) #2375
• [2ff183dd86] - doc: Update FIPS instructions in README.md (Michael Dawson) #2278
• [6483bc2e8f] - doc: clarify options for fs.watchFile() (Rich Trott) #2425
• [e76822f454] - doc: multiple documentation updates cherry picked from v0.12 (James M Snell) #2302
• [1738c9680b] - net: ensure Socket reported address is current (Ryan Graham) #2095
• [844d3f0e3e] - path: use '===' instead of '==' for comparison (Sam Stites) #2388
• [7118b8a882] - path: remove dead code in favor of unit tests (Nathan Woltman) #2282
• [34f2cfa806] - src: better error message on failed Buffer malloc (Karl Skomski) #2422
• [b196c1da3c] - src: fix memory leak in DLOpen (Karl Skomski) #2375
• [d1307b2995] - src: don't use fopen() in require() fast path (Ben Noordhuis) #2377
• [455ec570d1] - src: rename Buffer::Use() to Buffer::New() (Ben Noordhuis) #2352
• [fd63e1ce2b] - src: introduce internal Buffer::Copy() function (Ben Noordhuis) #2352
• [5586ceca13] - src: move internal functions out of node_buffer.h (Ben Noordhuis) #2352
• [bff9bcddb6] - src: plug memory leaks (Ben Noordhuis) #2352
• [ccf12df4f3] - (SEMVER-MINOR) src: add total_available_size to v8 statistics (Roman Klauke) #2348
• [194eeb841b] - test: drop Isolate::GetCurrent() from addon tests (Ben Noordhuis) #2427
• [46cdb2f6e2] - test: lint addon tests (Ben Noordhuis) #2427
• [850c794882] - test: refactor test-fs-watchfile.js (Rich Trott) #2393
• [a3160c0a33] - test: correct spelling of 'childProcess' (muddletoes) #2389
• [e51f90d747] - test: option to run a subset of tests (João Reis) #2260
• [cc46d3bca3] - test: clarify dropMembership() call (Rich Trott) #2062
• [0ee4df9c7a] - test: make listen-fd-cluster/server more robust (Sam Roberts) #1944
• [cf9ba81398] - test: address timing issues in simple http tests (Gireesh Punathil) #2294
• [cbb75c4f86] - tls: fix throughput issues after incorrect merge (Fedor Indutny) #2381
• [94b765f409] - tls: fix check for reused session (Fedor Indutny) #2312
• [e83a41ad65] - tls: introduce internal onticketkeycallback (Fedor Indutny) #2312
• [fb0f5d733f] - (SEMVER-MINOR) tools: run the tick processor without building v8 (Matt Loring) #2090
• [7606bdb897] - (SEMVER-MINOR) util: display constructor when inspecting objects (Christopher Monsanto) #1935

Crap. I just noticed all the changelog dates say the 18th (yesterday)... Building as the 19th.. hope it doesn't matter: https://jenkins-iojs.nodesource.com/job/iojs+release/76/

Slight issue with the Changelog from http://iojs.org/

@alexlamsl see #2448 :s