Audit commits not found on v4.x-staging

[MylesBorins]

I just did a run of branch diff and will put the results in a comment below. It appears there are 68 patch level commits that have landed on master that are not currently under LTS watch

branch-diff v4.x-staging master --exclude-label semver-major,semver-minor,lts-watch-v4.x,dont-land-on-v4.x --filter-release

We should likely audit these commits to see if we want to add them. I also think it would be a good idea for us to implement a new label to filter these results out in future audits. dont-land-on-v4.x?

[MylesBorins]

• [3470574cb6] - doc: clarify explanation of first stream section (Vitor Cortez) #4234
• [8182ec094d] - build: add option to select VS version (julien.waechter) #4645
• [d91646b9c7] - doc: rebuild LICENSE using tools/license-builder.sh (Rod Vagg) #4194
• [031b87d42d] - tools: add license-builder.sh to construct LICENSE (Rod Vagg) #4194
• [d1aabd6264] - test: fix style issues after eslint update (Michaël Zasso) nodejs/io.js#2286
• [ec8e0ae697] - tools: fix style issue after eslint update (Michaël Zasso) nodejs/io.js#2286
• [7ce0e04f44] - lib: fix style issues after eslint update (Michaël Zasso) nodejs/io.js#2286
• [4d5ee7a512] - tools: update eslint config (Michaël Zasso) nodejs/io.js#2286
• [2d441493a4] - tools: update eslint to v1.10.3 (Michaël Zasso) nodejs/io.js#2286
• [ed55169834] - gitignore: never ignore debug module (Michaël Zasso) #2286
• [3c70dc83ed] - doc: add CTC meeting minutes 2016-01-06 (Rod Vagg) #4667
• [55326f5488] - doc: add CTC meeting minutes 2015-12-16 (Rod Vagg) #4666
• [1a346840bf] - doc: add CTC meeting minutes 2015-12-02 (Rod Vagg) #4665
• [3ac5f8dcba] - doc: add CTC meeting minutes 2015-11-11 (Rod Vagg) #4664
• [db39625a44] - doc: add CTC meeting minutes 2015-11-04 (Rod Vagg) #4663
• [4383acd9f4] - build: fix and refactor VTune config in vcbuild.bat (Rod Vagg) #4192
• [be0b0b8cb9] - build: minor corrections in VTune configure text (Rod Vagg) #4192
• [7c60328002] - module: optimize js and json file i/o (Ben Noordhuis) #4575
• [038b636562] - module: avoid ArgumentsAdaptorTrampoline frame (Ben Noordhuis) #4575
• [83f8d98806] - module: cache stat() results more aggressively (Ben Noordhuis) #4575
• [809bf5e38c] - fs: change statSync to accessSync in realpathSync (Ben Noordhuis) #4575
• [05f17db8b2] - tools: fix warning in doc parsing (Shigeki Ohtsu) #4537
• [64e294adb2] - doc: fix heading level error in Buffer doc (Shigeki Ohtsu) #4537
• [7c8efeb6b4] - doc: fix website wg mislink (jona) #4357
• [4e155e8e2e] - test: write to tmp dir rather than fixture dir (Rich Trott) #4489
• [842e5a10ac] - test: don't assume a certain folder structure (Johan Bergström) #3325
• [8132b942a7] - test: make temp path customizable (Johan Bergström) #3325
• [11136bf334] - test: extend timeout in Debug mode (Rich Trott) #4431
• [fd551c3f29] - test: fix flaky cluster-disconnect-race (Brian White) #4457
• [79dc1d7635] - src: remove forwards for v8::GC*logueCallback (Ali Ijaz Sheikh) #4381
• [0b43c08f44] - util: pass on additional error() args (Brian White) #4279
• [cb0b4a6bc0] - test: add test for debugging one line files (cjihrig) #4298
• [2a60e2ad71] - module,src: do not wrap modules with -1 lineOffset (cjihrig) #4298
• [36ac3d642e] - deps: backport 8d6a228 from the v8's upstream (Fedor Indutny) #4259
• [da5cdc2207] - assert: accommodate ES6 classes that extend Error (Rich Trott) #4166
• [e2dec98837] - deps: upgrade to V8 4.7.80.25 (Ali Ijaz Sheikh) #4160
• [1e324d883e] - deps: backport bc2e393 from v8 upstream (evan.lucas) #4106
• [d9d050d396] - deps: cherry-pick 68e89fb from v8's upstream (Fedor Indutny) #4106
• [edfc8cde04] - deps: backport 1ee712a from V8 upstream (Julien Gilli) #4106
• [dc09bbe3ee] - test: fix test-repl-tab-complete after V8 upgrade (Ali Ijaz Sheikh) #4106
• [8a43a3d761] - deps: upgrade V8 to 4.7.80.24 (Ali Ijaz Sheikh) #4106
• [f88b1576e5] - deps: backport a7e50a5 from upstream v8 (Ben Noordhuis) nodejs/node-private#6
• [004778e6df] - deps: backport 6df9a1d from upstream v8 (Ben Noordhuis) nodejs/node-private#6
• [7e4f22ca13] - doc: update links to use https where possible (jpersson) #4054
• [b33e9da8f9] - debugger: introduce exec method for debugger (Jackson Tian)
• [70405d47c3] - deps: backport bc2e393 from v8 upstream (evan.lucas) #3792
• [81e32ccdfe] - deps: cherry-pick 68e89fb from v8's upstream (Fedor Indutny) #3779
• [2e38079ea4] - docs: fs - remove encoding list and link to buffer (fansworld-claudio)
• [6d6bc5d9d8] - test: mark http-pipeline-flood flaky (Rich Trott) #3616
• [be65f5f250] - tools: fix gyp to work on MacOSX without XCode (Shigeki Ohtsu) iojs/io.js#1325
• [1d65b99d5d] - tools: update gyp to b3cef02 (Imran Iqbal) #3487
• [494227bb03] - node: improve GetActiveRequests performance (Trevor Norris) #3375
• [680dda8023] - dns: remove nonexistant exports.ADNAME (Roman Reiss) #3051

[MylesBorins]

/cc @jasnell @rvagg @Fishrock123

[rvagg]

#2778 is an interesting case, it has landed-on-v4.x on it cause it was merged then I believe backed out. Is it time for a dont-land-on-v4.x flag to weed some of these out for good?

I think also it might be best to compare to v5.x rather than master, if it wasn't good enough for v5.x then it's not going to be good enough for v4.x.

[rvagg]

@iarna @Fishrock123 and @nodejs/v8: I've made a new dont-land-on-v4.x tag and am applying it to npm@3 and [email protected] PRs so we can filter them out.

[rvagg]

@iarna @Fishrock123 and @nodejs/v8: sorry, I meant to ask if you could please try and remember to tag new PRs with this label when we know they shouldn't touch v4.x.

[MylesBorins]

I've updated the list above filtering for dont-land-on-v4.x as well

[rvagg]

OK, I've done a heap of tagging and a bit of merging, if you grab the latest v5.x and run branch-diff against that you'll see the remaining list, which I have comments and questions on here (I think this covers it):

• [494227b] - node: improve GetActiveRequests performance (Trevor Norris) node: improve GetActiveRequests performance #3375 — interesting because it's a great perf improvement, not exposed via public API, but it is heavy-handed and not the API is not widely used so unlikely to be finding edge-cases very quickly. Perhaps it can be on a longer-term watch-list for LTS?
• [1d65b99] - tools: update gyp to b3cef02 (Imran Iqbal) tools: update gyp to b3cef02 #3487 — what's the status of AIX on v4.x? If we don't care about AIX for v4.x then we can probably ignore this one but it's a candidate for pulling in otherwise.
• [be65f5f] - tools: fix gyp to work on MacOSX without XCode (Shigeki Ohtsu) Upgrade to openssl-1.0.2a (Part1) #1325 - I don't understand why this didn't make it in, v4.x is on the same OpenSSL 1.0.2e as v5.x yet it was required for v4.x? @shigeki perhaps you can shed some light here?
• * [e302c33] - test: mark http-pipeline-flood flaky (Rich Trott) test: mark http-pipeline-flood flaky #3616 — we should try and have a similar test quality on LTS as we do on master, I find it difficult to keep track of what should be flaky and what shouldn't. Perhaps @Trott can help with cherry-picking?
  • [d09c447] - test: mark test-debug-no-context is flaky (Rich Trott) test: debug-no-context is flaky #4421 — same
• [81e32cc] - deps: cherry-pick 68e89fb from v8's upstream (Fedor Indutny) deps: cherry-pick 68e89fb from v8's upstream #3779 — see discussion in there, this may be something we want in LTS.
• [483a41c] - tls: copy client CAs and cert store on CertCb (Fedor Indutny) tls: copy client CAs and cert store on CertCb #3537 — too far outside my comfort zone to make a clear call on, see issue for discussion.
• [7a84fa6] - docs: fs - remove encoding list and link to buffer (fansworld-claudio) — see note there, one of two commits was landed, the other doesn't land cleanly, if author doesn't submit a PR we can just ignore IMO.
• [da5cdc2] - assert: accommodate ES6 classes that extend Error (Rich Trott) assert: accommodate ES6 classes that extend Error #4166 — I'd +1 this for LTS, I agree that it can be considered a bugfix
• [56bed48] - module,src: do not wrap modules with -1 lineOffset (cjihrig) Fix debugger and profiler functionality #4298 — confusing, see commentary in there
• The 6 commits in "tools: update eslint to v1.x" tools: update eslint to v1.x #2286 I guess are good to go, have we been doing these eslint changes in v4.x @thealphanerd?

[jasnell]

After having been delayed working on this by dealing with several fires this week, I'm hoping to be able to return attention to this later on today. @thealphanerd and @rvagg ... thank you both very much for continuing to press forward on it. I'm resetting my goal to have the 4.2.5 release proposal ready to go by end of the day Monday.

[MylesBorins]

@rvagg thanks for going through this!

Regarding linting changes, we have been backporting them for now. I imagine the number of rules being added are going to be minimal. The churn though will not. Keeping up with these changes will make future backporting substantially easier

[targos]

I cherry-picked the eslint update commits and fixed the merge conflicts: #4721

[Trott]

• [6d6bc5d9d8] - test: mark http-pipeline-flood flaky (Rich Trott) #3616 — we should try and have a similar test quality on LTS as we do on master, I find it difficult to keep track of what should be flaky and what shouldn't. Perhaps @Trott can help with cherry-picking?

This is a one-line change that was subsequently undone by d9734b7 (which fixed the test so it wasn't flaky anymore). I'd recommend cherry-picking them both (so that everything lands cleanly) and landing.

In fact, here's a PR that does exactly that: #4730

UPDATE: Er, actually, these are superseded by a third commit that's already in v4.x-staging. Cherry-picking both of these commits results in no change. So, no need to land them, I guess.

[Trott]

• [d09c447] - test: mark test-debug-no-context is flaky (Rich Trott) test: debug-no-context is flaky #4421 — same

This one cherry-picks cleanly to v4.x-staging. Here's a PR for good measure: #4732

[Trott]

• [da5cdc2] - assert: accommodate ES6 classes that extend Error (Rich Trott) assert: accommodate ES6 classes that extend Error #4166 — I'd +1 this for LTS, I agree that it can be considered a bugfix

Yup. If it's at all helpful, PR: #4734

I think that's the end of the stuff on the list that I was involved with. If there's something else I should look at, let me know.

[Trott]

@thealphanerd Is this issue still an active concern?

[MylesBorins]

Closing in lieu of a new thread to be opened soon