Jan 2017 Meeting

calendar_today January 19, 2017 confirmation_number 58 tickets discussed Source English

DevMeeting-2017-01-19

DevelopersMeeting20170119Japan

Date: 2017/01/19 (Wed) Time: 14:00- 19:00 (JST) Place: Money Forward inc. Headquarter Sign-up: https://ruby.connpass.com/event/47745/ Attendees: TBD Language: mostly Japanese (sorry for non native Japanese speakers)

Please add your favorite ticket numbers you want to ask to discuss.

  • NOTE
    • 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

  • NOTE: Write at least “ticket number/title/link” and your name (see example below). Explain details on the ticket. If you cannot attend the meeting, we appreciate a short summary because we can understand it more easily (long discussion is difficult to read, especially in a non-native language). Your motivation is also welcome.

About 2.5 timeframe

  • https://bugs.ruby-lang.org/projects/ruby-trunk/wiki/ReleaseEngineering25

Carry-over from previous meeting(s)

  • Postponed issues to avoid last-minute commits to 2.4:
    • [Feature #12180] switch id_table.c variant (shyouhei)
    • [Feature #12944] Change Kernel#warn to call Warning.warn (shyouhei)
      • related:
      • [Feature #13124] Should #puts convert to external encoding? (shyouhei)
    • [Feature #13017] Switch SipHash from SipHash24 to SipHash13 (shyouhei) I’m confident it’s OK to go to 2.5.
    • [Bug #12998] paragraph mode inconsistency between IO#each_line and String#each_line(nobu)
    • [Feature #8158] lightweight structure for loaded features index (shyouhei)
    • [Feature #12854] Proc#curry should return an instance of the class, not Proc (shyouhei)
  • [Feature #5481] Gemifying Ruby standard library (hsbt)
  • Waiting approval of Matz
  • [Feature #12901] Anonymous functions without scope lookup overhead (shyouhei)
  • [Feature #12906] do/end blocks work with ensure/rescue/else (shyouhei)
  • [Feature #12926] -l flag for line end processing should use chomp! instead of chop! (shyouhei)
  • [Feature #12912] An endless range (1..) (shyouhei)
  • [Feature #12933] Add Some and Optional (shyouhei)
  • [Feature #12931] Add support for Binding#instance_eval (shyouhei)
  • [Feature #12929] ternary should look ahead w/in a block (and not care about newlines) (shyouhei)
  • [Misc #12935] Webrick: Update HTTP Status codes, share them (shyouhei)
  • [Feature #12962] Feature Proposal: Extend ‘protected’ to support module friendship (shyouhei)
  • [Feature #12957] A more OO way to create lambda Procs (shyouhei)
  • [Feature #12966] net/ftp to include fxp support? (shyouhei) who’s going to handle this?
  • [Feature #12967] Add a default for RUBY_GC_HEAP_GROWTH_MAX_SLOTS out-of-the-box (shyouhei)
  • [Feature #12969] Allow optional parameter in String#strip and related (shyouhei)
  • [Feature #12968] Allow default value via block for Integer(), Float() and Rational() (shyouhei)
  • [Feature #12995] Conditional expression taking a receiver outside the condition (shyouhei)
  • [Bug #12998] paragraph mode inconsistency between IO#each_line and String#each_line (nobu)
  • [Feature #12996] Optimize Range#=== (shyouhei)
  • [Feature #10912] Add method(s) to IPAddr for determining whether an address is link local (shyouhei)
  • [Bug #13005] Inline rescue is inconsistent when rescuing NoMethodError (Dürst)
  • [Feature #13016] String#gsub(hash) (shyouhei)
  • [Bug #8241] If uri host-part has underscore ( ‘_’ ), ‘URI#parse’ raise ‘URI::InvalidURIError’ (shyouhei) go WHATWG or …?
  • [Bug #8826] BigDecimal#div and #quo different behavior and inconsistencies (shyouhei) status?>mrkn
  • [Feature #13026] Public singleton methods (shyouhei)
  • [Feature #8158] lightweight structure for loaded features index (shyouhei)
  • [Bug #13024] Confusing error message matching a non-ASCII string with ASCII-regex (shyouhei)
  • [Feature #13009] Implement fetch for Thread.current (shyouhei)
  • [Feature #13045] Passing a Hash with String keys as keyword arguments (shyouhei)
  • [Feature #13048] Better way to do Regexp.new(Regexp.escape(“some string”)) (shyouhei)
  • [Feature #9846] Regexp#to_regexp (shyouhei)
  • [Feature #13047] Use String literal instead of String#+ for multiline pretty-printing of multiline strings (shyouhei)

From attendees

  • Missed bugs to be assigned
    • [Bug #13111] Degraded performance for delegated methods through Forwardable module (shyouhei) assign who?
    • [Bug #13098] miniruby fails with SEGV on NetBSD/powerpc (shyouhei) look at the backtarce (shyouhei)
    • [Bug #13125] MRI has too much Qtrue : Qfalse; (shyouhei)
    • [Bug #13127] DRb `load’: connection closed (DRb::DRbConnError) when client exit’s from within a loop iterating over remote objects (shyouhei)
  • [Feature #13048] Better way to do Regexp.new(Regexp.escape(“some string”)) (shyouhei)
  • [Feature #8661] Add option to print backstrace in reverse order(stack frames first & error last) (shyouhei)
  • [Feature #13077] [PATCH] introduce String#fstring method (shyouhei)
  • [Feature #13083] {String Symbol}#match{?} with nil returns falsy as Regexp#match{?} (shyouhei)
  • [Feature #12508] Integer#mod_pow (shyouhei)
  • [Feature #13067] TrueClass,FalseClass to provide === to match truthy/falsy values. (shyouhei)
  • [Feature #8661] Add option to print backstrace in reverse order(stack frames first & error last) (shyouhei)
  • [Misc #13072] Current state of date standard library (shyouhei)
    • related:
    • [Bug #12981] Date.parse raises an Argument error under a specific condition (shyouhei) assign who?
    • [Bug #13101] Date#rfc2822 and Time#rfc2822 don’t return the same format
  • [Bug #13064] Inconsistent behavior with next inside begin/end across different implementations. (shyouhei) spec or…?
  • [Feature #13077] [PATCH] introduce String#fstring method (shyouhei)
  • [Feature #13083] {String Symbol}#match{?} with nil returns falsy as Regexp#match{?} (shyouhei)
  • [Feature #13097] Deprecate Socket.gethostbyaddr and Socket.gethostbyname (shyouhei)
  • [Feature #13067] TrueClass,FalseClass to provide === to match truthy/falsy values. (shyouhei)
  • [Bug #13104] math.rb affects Rational literals (nobu)
  • [Bug #13105] String#to_f and String#to_r don’t stop successive underscores (nobu)
  • [Feature #13109] using in refinements is required to be physically placed before the refined method call (shyouhei)
  • [Bug #12705] yielding args to a lambda uses block/proc rather than lambda/method semantics (shyouhei) status?
  • [Bug #9569] SecureRandom should try /dev/urandom first (shyouhei) I now think it’s OK
  • [Feature #13095] [PATCH] io.c (rb_f_syscall): remove deprecation notice (shyouhei)
  • [Feature #13122] Special syntax for Hash#default_proc (shyouhei)

From non-attendees

Write your name and your interest (what do you want to ask and to whom?) please.

  • In-tree copy of ruby spec (see http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-core/78985) (duerst)
  • example: [Feature #10917] Add GC.stat[:total_time] when GC profiling enabled (ko1)

Log

Date: 2017/01/19 (Wed)

Time: 14:00- 19:00 (JST)

Place: Money Forward inc. Headquarter

Sign-up: https://ruby.connpass.com/event/47745/

log edit: https://docs.google.com/document/d/1ZKk-vxoYkq8b2H4ml2z4NhoHsi3GdZqhNXNgpB2vTv8/edit

log: https://docs.google.com/document/d/1ZKk-vxoYkq8b2H4ml2z4NhoHsi3GdZqhNXNgpB2vTv8/pub

Next meeting

  • 2/22 at MoneyForward HQ
  • Another place?

About 2.5 timeframe

https://bugs.ruby-lang.org/projects/ruby-trunk/wiki/ReleaseEngineering25

  • 合宿

  • Want to do
  • 1Q - 2Q
  • May through July?

Current status of Ruby 2.4

  • Mentainer not fixed yet.

  • RA, nagachika, and usa should be consulted.

  • March is too late.

  • akr: 2.1’s security fix expires in this Feb.

Carry-over from previous meeting(s)

[Feature #12180] switch id_table.c variant (shyouhei)

  • ko1: funny-falcon’s open addressing hash is beating mine.
  • matz: why not just commit?
  • ko1: I want to let him merge himself.
  • akr: isn’t it too big for a new comer?
  • ko1: I see. I’ll do.

[Feature #12944] Change Kernel#warn to call Warning.warn (shyouhei)

  • nobu: what spec?
  • shyouhei: give up passing Arrays to it.
  • akr: compatibility concern?
  • matz: why not just try it.

[Feature #13124] Should #puts convert to external encoding? (shyouhei)

  • shyouhei: what’s this?
  • akr: does it depend on internal encoding?
  • naruse: yes.
  • nobu: the String knows its own encoding.  why not honor that?
  • naruse: current idea is that we don’t automatically encode things.
  • akr: I think it’s OK for particular IO that needs encoding, it’s convenient to have such feature.
  • naruse: OK, I’ll respond.

[Feature #13017] Switch SipHash from SipHash24 to SipHash13 (shyouhei)

  • shyouhei: I’m confident it’s OK to go to 2.5.
  • nobu: SipHash24 was too strong.
  • hsbt: who is going to apply?
  • shyouhei: I will
  • matz: go ahead.

[Bug #12998] paragraph mode inconsistency between IO#each_line and String#each_line(nobu)

  • ko1: ideal behaviour?
  • nobu: I think IO’s is better.
  • matz: (nod)
  • nobu: I already merged this.

[Feature #8158] lightweight structure for loaded features index (shyouhei)

  • ko1: nobody but nobu can say if it’s OK.
  • nobu: I’ll take a look.
  • hsbt: why not let funny-falcon merge this?
  • matz: I take the idea.

[Feature #12854] Proc#curry should return an instance of the class, not Proc (shyouhei)

  • ko1: general idea?
  • matz: basically we honor sub-classing but Array has exceptional situations.
  • shyouhei: seems there are applications for this feature?
  • matz: seems they are different proposals than applications of #curry.
  • matz: I’ll ask for use cases.

[Feature #5481] Gemifying Ruby standard library (hsbt)

  • hsbt: Waiting approval of Matz
  • matz: rgr.
  • naruse: isn’t it better to convert one by one?
  • akr: there can be distinct discussions for each libraries.

  • nobu: there are 3 kinds of libraries:

  • standard libray,
  • default gem, which are inside of source tree, comes with gemspec

  • bundled gem, which are out of tree, installed automatically on make install

  • hsbt: I’m talking about default gem.
  • hsbt: practically what happens would be to create gemspec for each libraries.
  • akr: I see concerns of dependency hell.
  • hsbt: it’s OK we create separate ML threads for each library to discuss.

[Feature #12901] Anonymous functions without scope lookup overhead (shyouhei)

  • ko1: matz said we should automatically optimize procs when possible, not explicitly stating like this.
  • matz: (nod)
  • akr: is that possible? it’s difficult to detect.
  • ko1: lets make it possible.  say, by restricting use of Proc#binding.
  • ko1: one problem on restricting binding access is that should kill debuggers to inspect running methods.  But I think that’s acceptable.

  • ko1:

  • a=b=c=nil; iter do; use(a); use(b); eval(“use(c)”); end # NG
  • a=b=c=nil; iter do; binding; use(a); use(b); eval(“use(c)”); end # OK
  • x = 1; func{ a a+1}
  • matz will respond, ko1 to create a new ticket that refer this one.

[Feature #12906] do/end blocks work with ensure/rescue/else (shyouhei)

  • shyouhei: OP is wise because he do not talk about { … }
  • matz: I don’t favor too small rescue blocks.
  • naruse: I write this occasionaly.
  • akira: and when I do so, I have to re-indent everything.
  • naruse: HTTP or RDB networking codes tends to do error handling.  There are chances for its usage.
  • akr: generally speaking it is good to reduce indent.
  • matz: Let me think about it. -> OK

[Feature #12926] -l flag for line end processing should use chomp! instead of chop! (shyouhei)

  • nobu: perl behaves like chomp.
  • akr: isn’t it a bug?
  • matz: that’s because chomp is newer than -l.
  • akr: chomp is ruby1.1+, -l already exists in ruby 0.95
  • matz: OK.

[Feature #12912] An endless range (1..) (shyouhei)

  • shyouhei: what about opossite side …1?
  • mrkn: I think what is requested is about sequence, not range.
  • akr: Range is defined over #succ, so 1.. is possible.
  • nobu: 1..*
  • akr: 1.step works. isn’t it enough?
  • shyouhei: is iteration the only usage?
  • naruse: array index.
  • mrkn: index for multi-dimensional array can be a possible application. Python’s numpy has such feature.
  • akr: I doubt if we need a dedicated syntax for this.
  • nobu: (1..) seems visually odd.
  • akr: for now, use step and drop.

[Feature #12933] Add Some and Optional (shyouhei)

  • nobu: start as a gem.
  • matz: agree.

[Feature #12931] Add support for Binding#instance_eval (shyouhei)

  • shyouhei: what is needed ultimately is to pass a block to Binding#eval
  • ko1: I don’t like it.  We have to reconstruct everything if we do this.
  • nobu: I think it’s not faster, if not impossible.
  • nobu will reject.

[Feature #12929] ternary should look ahead w/in a block (and not care about newlines) (shyouhei)

  • naruse: perl people tends to write it.
  • nobu: it’s possible if parens are there, but I don’t want to look ahead.
  • shyouhei: Martin says we already look-ahead for periods.
  • matz: Python ignores newlines inside of parens.

[Misc #12935] Webrick: Update HTTP Status codes, share them (shyouhei)

  • nobu: is it Webrick?
  • akr: Webrick has this code but no one wants to require the whole.
  • akira: nobody looks at anywhere but Rack.

[Feature #12962] Feature Proposal: Extend ‘protected’ to support module friendship (shyouhei)

  • akr: use of protected is discouraged.
  • hsbt: deprecate protected?
  • akira: possible now. Rails now deletes most use of protected.
  • naruse: I think it’s RuboCop’s duty to warn such thing.
  • akr: it’s a good thing anyway to update rdoc to discourage usages.

[Feature #12957] A more OO way to create lambda Procs (shyouhei)

  • akr: I don’t understand the OP’s intension.
  • shyouhei: I think they want a subclass of lambda.  that’s not possible today.
  • akr: it is possible. complicated though. ->  commented.

[Feature #12966] net/ftp to include fxp support? (shyouhei) who’s going to handle this?

  • shugo is handling.

[Feature #12967] Add a default for RUBY_GC_HEAP_GROWTH_MAX_SLOTS out-of-the-box (shyouhei)

  • ko1: I’m not sure if this parameter is a sane default.
  • ko1: also, I think there should be a comprehensive approach to parameterize GC, not individually like this.
  • naruse: I think regexp is the best for this purpose.
  • naruse: I don’t like the tr-like argument string.

[Feature #13016] String#gsub(hash) (shyouhei)

  • akr: what’s wrong with union?
  • shyouhei: that’s not what I want.
  • shyouhei: All what I want to do is to table lookup. Not regular expression.
  • akr: performance concern.
  • shyouhei: understand that point.
  • naruse: you have to sort by length before union. Otherwise カ can match before ガ.

[Bug #8241] If uri host-part has underscore ( ‘_’ ), ‘URI#parse’ raise ‘URI::InvalidURIError’ (shyouhei)

  • naruse: I think this is done.

[Bug #8826] BigDecimal#div and #quo different behavior and inconsistencies (shyouhei) status?>mrkn

  • mkrn: I’ll start to fix it as soon as possible.

[Feature #13048] Better way to do Regexp.new(Regexp.escape(“some string”)) (shyouhei)

  • naruse: quick look at the github result seems every usage of Regexp.new(Regexp.escape()) are bad.
  • akr: Regexp.exact() ?

[Bug #13111] Degraded performance for delegated methods through Forwardable module (shyouhei) assign who?

  • nobu: I did this already I think.

[Bug #13098] miniruby fails with SEGV on NetBSD/powerpc (shyouhei) look at the backtarce (shyouhei)

  • nobu: will take a look.

[Bug #13125] MRI has too much Qtrue : Qfalse; (shyouhei)

  • nobu: nobody is against the needs of macro but the name?
  • shyouhei: will update the ticket.

[Bug #13127] DRb `load’: connection closed (DRb::DRbConnError) when client exit’s from within a loop iterating over remote objects (shyouhei)

  • assign seki

[Feature #13067] TrueClass,FalseClass to provide === to match truthy/falsy values. (shyouhei)

  • matz: I give up.

[Bug #13064] Inconsistent behavior with next inside begin/end across different implementations. (shyouhei) spec or…?

  • ko1: it’s 1 to me.
  • matz: fail early is good but is it a syntax error?
  • ko1:  it is.  Because you can’t place next here statically.
  • akr: should the JIS/ISO specify as such?

[Bug #13104] math.rb affects Rational literals (nobu)

  • matz: it’s ugly, but mathn is ugly by nature.

[Bug #9569] SecureRandom should try /dev/urandom first (shyouhei) I now think it’s OK

  • shyouhei: I will

  • rename Random.raw_seed to Random.urandom
  • try to use that from SecureRandom

  • if it fails, fallback to OpenSSL.

  • [Bug #13135] Regexp.last_match returns nil with s.rindex(//)

  • everyone: isn’t it a bug?
  • ko1: assign shugo.
  • akr: it seems to be an optimization.
  • akr: regular expression should set $~.  Optimization should be done against empty string.