Aug 2018 Meeting

calendar_today August 09, 2018 confirmation_number 23 tickets discussed Source English

DevMeeting-2018-08-09

Please comment your favorite ticket numbers you want to ask to discuss with your SHORT comment or summary. (your summary/comment will help us because we don’t need to read all of ticket comments)

DO NOT discuss then on this ticket, please.

NOTES

  • Dev meeting IS NOT a decision making place. All decisions should be done at the bug tracker.
  • Dev meeting is a place we can ask Matz, nobu, nurse and other developers directly.
  • Matz is a very busy person. Take this opportunity to ask him. If you can not attend, other attendees can ask instead of you (if attendees can understand your issue).
  • We will write a log about discussion to a file or to each ticket in English.
  • All activities are best-effort (keep in mind that most of us are volunteer developers).
  • The date, time and place is scheduled according to when/where we can reserve Matz’s time.

Agenda

Next dev-meeting

About 2.6 timeframe

Carry-over from previous meeting(s)

  • [Bug #14887] Array#delete_if does not use #delete (shyouhei)
    • Is it by design, or a bug?
  • [Feature #13050] Readline: expose rl_completion_quote_character variable (nobu)
  • [Feature #14850] Add official API for setting timezone on Time (nobu)
  • [Feature #14869] Proposal to add Hash#=== (nobu)
  • [Feature #14916] Proposal to add Array#=== (aycabta)
  • [Feature #14877] Calculate age in Date class (nobu)
  • [Bug #14908] Enumerator::Lazy creates unnecessary Array objects. (nobu)

From Attendees

(will be edited later) (if you have a write access permission, please list directly)

  • [Feature #5446] at_fork callback API (eregon) Can we get approval for introducing this? It is solving a real-world issue, in a more reliable way than the various hacks in most Ruby servers.
  • [Feature #11258] add ‘x’ mode character for O_EXCL (kazu)
  • [Misc #14907] io.c: do not close inherited FDs by default (akr)
  • [Feature #14973] Proposal of percent literal to expand Hash (aycabta)

From non-attendees

  • [Feature #14717] thread: allow disabling preempt (normalperson)

  • [Feature #11076] Enumerable method count_by (baweaver) As mentioned in the discussion, Nobu was kind enough to update this to work with current master for Ruby: https://github.com/ruby/ruby/compare/trunk…nobu:feature/11076-Enumerable%23count_by

    As the code is already written I believe it would be an ideal target for 2.6 as a quick win on an often requested feature.

  • [Feature #14954] Add :wait option to RubyVM::MJIT.pause (k0kubun)
    • Since it’s not released even in preview2 yet, is it okay to change the behavior?
    • Is there any objection for having the option or the option name?
  • (baweaver)
    • [Feature #14869] Proposal to add Hash#=== (nobu)
    • [Feature #14916] Proposal to add Array#=== (aycabta)

    • [Feature #14912] Introduce pattern matching syntax

      I believe that these three items are related, and personally I could do a lot of very interesting things with Array#=== and Hash#=== that would get us exceptionally close to features discussed in [Feature #14709]

      This to say I would second the inclusion of the first two items as they relate to the third and fourth. It would give Ruby an exceptional amount of power.

  • [Feature #14759] [PATCH] set M_ARENA_MAX for glibc malloc (sam.saffron) I would also very much like to see it backported 2.4 and 2.5. It is of enormous value to the ecosystem
  • [Feature #14736] Thread selector for flexible cooperative fiber based concurrency (ioquatix)
  • [Feature #14739] Improve fiber yield/resume performance (ioquatix)
  • [Bug #14968] io.c: make all pipes nonblocking by default (normalperson)

  • [Misc #14937] timer-thread elimination depends on [Bug #14968] (normalperson)

  • [Feature #14944] Support optional inherit argument for Module#method_defined? (jeremyevans0)

(will be edited later) (if you have a write access, please list directly)

Log

Next dev-meeting

  • 9/13

  • @ Cookpad

About 2.6 timeframe

  • Naruse: nothing to do right now.
  • Mame: feature freeze?
  • Naruse: expected in Oct. maybe.

Carry-over from previous meeting(s)

  • [Bug #14887] Array#delete_if does not use #delete (shyouhei)

  • Is it by design, or a bug?
  • Matz: delete is memory inefficient.

  • Mame: if you can override #delete, why not also #delete_if.

  • [Feature #13050] Readline: expose rl_completion_quote_character variable (nobu)

  • Go!!!

From Attendees

  • [Feature #5446] at_fork callback API (eregon) Can we get approval for introducing this? It is solving a real-world issue, in a more reliable way than the various hacks in most Ruby servers.

  • Mrkn: I wanted to use this in pycall to call PyOS_AfterFork_Child, and so on from C-layer, but I noticed that I can use pthread_atfork directly.
  • Naruse: Ruby’s fork is unsafe operation. Application programmers should understand the fact and carefully handle related operations (for example freeing resources).
  • Akr: I guess at_fork hook would be problematic with multi-thread and/or multi-guild.  But making it Ruby’s standard feature impress this feature usable universally including multi-thread/guild. So, I’m negative for this proposal.

  • usa: Furthermore, I consider that fork must be destroyed.

  • [Feature #11258] add ‘x’ mode character for O_EXCL (kazu)

  • Matz accepted already.

  • [Misc #14907] io.c: do not close inherited FDs by default (akr)

  • Nobody is against changing defaults

  • [Feature #14850] Add official API for setting timezone on Time (nobu)

  • Akr: localtime/timelocal analogous variant methods should suffice

  • Timezone object shall define localtime/timelocal/gmtime/timegm

  • Matz: is there any reason we need this?

  • Matz: What is actual API?

  • [Feature #14869] Proposal to add Hash#=== (nobu)

  • After [Feature #14912]

  • [Feature #14877] Calculate age in Date class (nobu)

  • Matz: I don’t like the idea to implicitly use today.

  • [Feature #14973] Proposal of percent literal to expand Hash (aycabta)

  • Matz: percent literal…

  • [Feature #13618] [PATCH] auto fiber schedule for rb_wait_for_single_fd and rb_waitpid (ko1)

  • Naming.

  • X IoThread.new{}
  • X GreenThread.new{}
  • X Thread::Green.new{} #124 “most likely candidates”(vote: hsbt)
  • X Thread.green{} #124 “most likely candidates”
  • X Thread.create(scheduler: …, args: …) #124 “most likely candidates”
  • Thread::Coop.new{}
  • Thread::Cooperative.new{} # (usa)
  • Thread::Nonpreemptive.new{} # (usa; these 2 names are very long, then they are good ^^)
  • NonpreemptiveThread.new {} # (mrkn)
  • X NPThread.new {} # (mrkn)

  • X Thread::Light (ko1)

  • Matz: “Thread.green” and “Thread.create” are bad because we should distinguish auto fiber and Thread.
  • Matz: “Thread::Green” is bad because it means implementation.

From non-attendees

  • [Feature #14717] thread: allow disabling preempt (normalperson)

  • Closed already

  • [Feature #11076] Enumerable method count_by (baweaver) As mentioned in the discussion, Nobu was kind enough to update this to work with current master for Ruby: https://github.com/ruby/ruby/compare/trunk…nobu:feature/11076-Enumerable%23count_by As the code is already written I believe it would be an ideal target for 2.6 as a quick win on an often requested feature.

  • Matz: I thought #count_by would be something different.
  • Mame: There is #count. #count_by is very different from that.
  • K-tsj: Also, #count takes a block.
  • Matz: I’m not against the feature itself.  Also, do we need a block-less counterpart?

  • Naruse: It looks similar to Presto’s histogram function

  • [Feature #14954] Add :wait option to RubyVM::MJIT.pause (k0kubun)

  • Since it’s not released even in preview2 yet, is it okay to change the behavior?

  • Is there any objection for having the option or the option name?

  • Not. At. All.
  • Usa: Don’t ask for permission, beg for forgiveness.

  • Naruse: RubyVM::MJIT is your namespace.

  • (baweaver)

  • [Feature #14869] Proposal to add Hash#=== (nobu)
  • [Feature #14916] Proposal to add Array#=== (aycabta)
  • [Feature #14912] Introduce pattern matching syntax I believe that these three items are related, and personally I could do a lot of very interesting things with Array#=== and Hash#=== that would get us exceptionally close to features discussed in [Feature #14709] This to say I would second the inclusion of the first two items as they relate to the third and fourth. It would give Ruby an exceptional amount of power.

  • Matz: we should discuss these after pattern match is introduced.

  • [Feature #14759] [PATCH] set M_ARENA_MAX for glibc malloc (sam.saffron) I would also very much like to see it backported 2.4 and 2.5. It is of enormous value to the ecosystem

  • Usa: The decision should be made by Linux port maintainer
  • Naruse: up to kosaki.

  • Mame: Once kosaki said “If we merge this, we should decide when we remove this”.

  • [Feature #14944] Support optional inherit argument for Module#method_defined? (jeremyevans0)

  • Usa: I want this too.
  • Mame: why?
  • Usa: maybe when I want to develop.

  • Matz: not that negative.  Rather I think it’s nice to have.

  • [Feature #14975] String#append without changing receiver’s encoding (ioquatix)

  • Naruse: I think it’s OK to have this kind of operation(s), but not for #«

Carry over:

  • [Bug #14908] Enumerator::Lazy creates unnecessary Array objects. (nobu)
  • [Feature #14736] Thread selector for flexible cooperative fiber based concurrency (ioquatix)

  • [Feature #14739] Improve fiber yield/resume performance (ioquatix)

  • Ko1: will discuss them in person

  • [Bug #14968] io.c: make all pipes nonblocking by default (normalperson)
  • [Misc #14937] timer-thread elimination depends on Bug #14968