Skip to content

Latest commit

 

History

History
95 lines (72 loc) · 5.33 KB

2017-04-24.md

File metadata and controls

95 lines (72 loc) · 5.33 KB
Raw

Node.js LTS meeting 24 April 2017

Present

  • Michael Dawson (@mhdawson)
  • Myles Borins (@MylesBorins)
  • James Snell (@jasnell)
  • Jeremiah Senkpiel (@fishrock123)
  • Sam Roberts (@sam-github)
  • Gibson Fahnestock (@gibfahn)

Agenda

nodejs/LTS

  • Limit push access to LTS staging branches to backport team #199
  • Node-v8 LTS support end on 2019-12-31 to meet EOSL of openssl-1.0.2 #186
  • Backporting test fixes to maintenance branches #205
  • Inspector - retain support for --inspect --debug-brk nodejs/node#12364 (related: nodejs/node#12615)
  • Plan for triaging semver-minor backports in LTS meetings #204
  • Potential Semver Minor Backports #177

Minutes

Limit push access to LTS staging branches to backport team #199

  • Myles: knowing that only people with backporting auth allows more trust, as-is, myles has to audit the -staging branch before starting to land commits, which is time consuming.
  • Myles: also, everyone should turn on signed commits
  • Myles: doesn’t think 7.x-staging (current) needs this kind of protection, for now
  • Gibson: +1
  • Sam: +1
  • Michael: has only pushed to -staging when asked, would this be an issue?
  • Myles: would prefer to pick peoples commits to staging
  • James: +1
  • Myles: mark ctc-review and give it a few days to ensure no objections?
  • James: good idea

Node-v8 LTS support end on 2019-12-31 to meet EOSL of openssl-1.0.2 #186

  • Gibson: would shorten by 3 months
  • Gibson: objections?
  • All: … (no objections)
  • Myles: what is next version of OpenSSL with long support
  • Michael: 1.2?
  • Myles: concerned that there is not a more proactive approach to planning when we upgrade openssl
  • Michael: lets wait until 8.0, then start talking about it

Backporting test fixes to maintenance branches #205

  • Myles: there are already 2 test commits in that fix very broken behaviour
  • Myles: concerned about a slippery slope, and that there isn’t a clear policy, and that one test may pull more tests in
  • Myles: also concerned that anybody (not just LTS members) feel they have equal access to request backport, and if anybody can request backports, will there be too many?
  • James: active, we backport things we think can go, maintainance, we only backport when people request (and after considering the request)
  • Michael: feels it should be request driven, and based
  • Myles: what about semver-minor? Are we going to allow minors on things in maintenance?
  • Sam: I’m OK, if its high value, and VERY low risk
  • James: we’ve done this on a case-by-case basis in the past (0.10, 0.12)
  • Myles: OK, but we should change the process to express that people may request backports to the maintenance branch, and LTS WG will consider
  • Jeremiah: we should always backport tests along with features/fixes
  • Myles: need to add wording to that non-critical changes can land with LTS WG consensus, hopefully without a full meeting
  • Sam: thinks the test problem under consideration is critical…
  • Myles: wants critical to be high bar, something that EVERY user should update to get the fix for
  • Michael: willing to take a shot at a new language
  • Myles: “will land fixes into maintenance as considered necessary for the stability of the maintenance release” - but we are much less proactive in landing things. Stability is the most important value.
  • ... Much description
  • Myles: this specific test can land, if there are no objections, but we should change the docs and be very conservative
  • ... we all seem to want to be conservative in accepting changes into maintainence releases

Inspector - retain support for --inspect --debug-brk nodejs/node#12364 (related: nodejs/node#12615)

  • [2017-04-24 14:45] If you read me. I say remove --inspect-brk, it won't be adopted
  • [2017-04-24 14:46] refack: you listing to the LTS meeting? I'm OK with removing --inspect-brk as well.
  • Gibson: thinks we should backport --inspect-brk
  • Myles: node debug already aliased, why not alias --debug-brk?
  • Gibson: if we are going to use --inspect-brk in the future, we should backport it, if we are going to instead just use --debug-brk in the future, we should bring it back on master. Conversation about what to do should happen in a new issue.
  • [2017-04-24 15:02] octetcloud: new meta issue nodejs/node#12630

Plan for triaging semver-minor backports in LTS meetings #204

  • Gibson: can we agree on a async review process? If so, we can do triaging online.
  • ... no objections in theory

Potential Semver Minor Backports #177

  • Skipped this time as the list wasn't filtered.