Oct 2019 Meeting

calendar_today October 17, 2019 confirmation_number 21 tickets discussed Source English

DevMeeting-2019-10-17

Please comment on 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 the ticket comments)

DO NOT discuss then on this ticket, please.

Date: 10/17 13:00-17:00 Place and Sign-up: (See the Log) log: https://docs.google.com/document/d/1-IC98nOPY_FZnne2PXILYAaxt6PqHfhphAI908K3vD8

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 the 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 are scheduled according to when/where we can reserve Matz’s time.

Agenda

Next dev-meeting

About 2.7 timeframe

Check security tickets

Discussion

Please comment on your favorite ticket we need to discuss with the following format.

* [Ticket ref] Ticket title (your name)
* your comment why you want to put this ticket here if you want to add.

Your comment is very important if you are no attendee because we can not ask why you want to discuss it.

Example:

* [Feature #14609] `Kernel#p` without args shows the receiver (ko1)
* I feel this feature is very useful and some people say :+1: so let discuss this feature.

We don’t guarantee to put tickets in the agenda if the comment violates the format (because it is hard to copy&paste).

A short summary of a ticket is strongly recommended. We cannot read all discussion of the ticket in a limited time. A proposal is often changed during the discussion, so it is very helpful to summarize the latest/current proposal, post it as a comment in the ticket, and write a link to the comment.

Log

https://bugs.ruby-lang.org/issues/16232 Venue 10/17 (Thu) 13:00-17:00 @ pixiv Inc. (6F) up to 6F, and ask someone to call usa-san in Studio Attendees Add your name (or ask a committer to invite you) matz (remote) mame (remote) usa (venue staff) znz (remote) nobu (remote) aycabta osyo (invited by aycabta) a_matsuda mrkn (remote) akr Soutaro knu Absent: Martin Dürst (trip to Unicode Conference) Shyouhei (is sick)

Next Date 11/12 (Tue) 13:00-17:00 @ full remote Review all 2.7 new features 11/28 (Thu) 13:00-17:00 @ pixiv 6F studio Monthly meeting; it would be the last change for 2.7 new features Announce About 2.7 timeframe preview 2 naruse: within a few days, maybe Next Stable releases timeframe usa: maybe the 2nd week of December? mame: need to ask nagachika-san Check security tickets Topics [Feature #16029] Expose fstring related APIs to C-extensions (byroot) Apparently the current implementation is complicated to expose, but no details were provided. Is there something I or others could change to the implementation so that it could be exposed? Discussion: nobu: does it make speed up? naruse: yes see also https://twitter.com/search?q=from%3Afrsyuki until%3A2017-05-04&src=typed_query mame: will the strings be GC-ed? nobu: yes ko1: it’s a little dangerous nobu: the dangerousness is not of fstring but of shared string ko1: exposing these APIs may introduce many bugs on corner cases, I guess [Feature #16038] Provide a public WeakMap that compares by equality rather than by identity (byroot) Matz asked for real world use-case, which I did provide. In short it’s useful for implementing value objects deduplication, just like how the fstring table works, but for user defined, more complex types. Discussion: mame: the use-case is for memorize usa: it’s reasonable ko1: is GC-scheduled objects still be visible, isn’t it? nobu: lazy sweep checks it ko1: is finalizer OK? … ah, it may be OK mame: is it OK that the referenced object is lazy sweeping? nobu & matz: there is no such case by design ko1: is compaction safe? only Aaron knows :) mame: the proposal requests an interface to use equality based map mame: introduce a new keyword argument for the flag? matz: I don’t disagree the feature itself ko1: for implementation, we have to make a new compare function for st? nobu: maybe ko1: and make a new check for the existance… nobu: maybe akr: I don’t like the API, compare_by_equality, because it introduces extra complexity: When it is called with non-empty WeakMap, it may remove some elements. mame: I agree akr. there might be much better API akr: I think that this proposal expects an ideal GC. The real application may behave badly if GC is invoked too frequently or too rare. usa: at all, should we implement this API in 2.7 or not? matz: we don’t have to be rough-and-ready in this case [Bug #10314] Default argument lookup fails in Ruby 2.2 for circular shadowed variable names (jeremyevans) Do we want to change def foo(bar=bar) from warning to SyntaxError using the patch? Discussion: nobu: we can make this an error now matz: go ahead ko1: method body? mame: should be in another ticket. should warn before to be an error ko1: matz, is the change ok? matz: ok [Bug #11055] autoload resets private_constant (jeremyevans) Do we want to copy constant visibility information across the autoload using the patch? Discussion: usa: is this bug? matz: maybe. but if a library does not recognize the constant is private, how autoload behave? nobu?: Current implementation is to delete the constant from the entry, and add new one, then private flag is cleared matz: how to fix? nobu: I’ll look this in a few days. [Bug #13249] Access modifiers don’t have an effect inside class methods in Ruby >= 2.3 (jeremyevans) Do we want to add a warning for misuse of method visibility methods using the patch? Discussion: already fixed. [Bug #15267] File.basename + File.extname does not restore the original name (jeremyevans) Do we want to make them restore the original name using the patch? Discussion: name = ‘file.’ File.basename(name, ‘.*’) #=> “name” File.extname(name) #=> “”

usa: should fix File.extname, not File.basename, because of their usecases, if we do File.extname(“foo.”) will be “.” matz: ok [Feature #16245] Add interfaces to count and measure size all IMEMO objects (Sam Saffron) Would love to see this included in 2.7, refined proposal does not add any new APIs. Discussion: RubyVM.stat { :global_method_state=>143, :global_constant_state=>1369, :class_serial=>8768, :imemo_ment_count, :imemo_iseq_count, :imemo_env_count, :imemo_tmpbuf_count, :imemo_ast_count, :imemo_ment_size, :imemo_iseq_size, :imemo_env_size, :imemo_tmpbuf_size, :imemo_ast_size }

ko1: I’ve already responded. [Feature #13683] Add strict Enumerable#single (rafaelfranca) Matz didn’t like the single name but we have a proposal for #only Discussion: [1].only #=> 1 [1, 2].only #=> ArgumentError

ko1: pattern match may be helpful for this use case? case [1] ; in (x); p x; end #=> [1] case [1, 2]; in (x); p x; end #=> [1, 2] case [1] ; in [x]; p x; end #=> 1 case [1, 2]; in [x]; p x; end #=> NoMatchingPatternError

a_matsuda: This is actually wanted in Rails applications. It named “#pick”. matz: I hate #only. #first! is not good because it does not represent the behavior naruse: the name need to include both checking case [1, 2] in [x] p x end #=> t.rb:3:in `<main>’: [1, 2] (NoMatchingPatternError)

[Bug #16143] BOM UTF-8 is not removed after rewind (kou) Can we commit this? I’ve reviewed. I think that this is ready to merge. Discussion: usa: rewind doesn’t sound like encoding aware. IO#rewind is really low level operation, we don’t call it usually. CSV#rewind should handle BOM itself, not IO#rewind. Should provide a method to know BOM is read? Adding an option io.rewind(respect_bom: true) can be an option??? rewind(to_after_bom: true) rewind(skip_bom: true) naruse: If a code rewind and write BOM, this change will break such code. Let CSV#rewind know if it need to handle bom: test if the input is unicode and skip the bom??? Add Encoding#unicode? and IO#skip_bom => rewind(skip_bom: true) [Feature #15822] Add Hash#except (zverok) The method is really useful, even with ActiveSupport-less codebases I constantly tend to redefine it with core ext or refinements. Hash#slice was merged in 2.5; it was initially discussed in #8499 alongside the except, but from the discussion, it is not obvious why #except was “lost” :( Discussion: ActiveSupport: https://api.rubyonrails.org/classes/Hash.html#method-i-except

File activesupport/lib/active_support/core_ext/hash/except.rb, line 12

def except(keys) slice(self.keys - keys) end

knu: Use cases: super(*options.expect(:foo)), json.except(“metadata”), params.except(:id), etc. naruse: Hash#slice was introduced in the context of #13563, not #8499. In #13563, Hash#except is not discussed. That’s the reason why #except was “lost”. matz: Want to see the actual use case. And I don’t like the name “except”. Q. What about extending select/reject to alternatively take keys instead of a block? A. (by knu) hash.select()/hash.reject() currently returns an Enumerator, so hash.select(keys) wouldn’t work as intended (returning a slice of hash) when keys were empty and that’d be confusing and error prone. [Feature #16131] Remove $SAFE, taint and trust (jeremyevans) Does anyone have time to review the patch (extensive changes, mostly code deletion)? How do we want to handle included libraries with separate upstreams that want to be compatible with older Ruby versions (bundler, rubygems, etc.)? Discussion: akr: Adding new mechanism to make warnings for transitions would make sense (the warnings are not printed with -v.) No recommendation from ruby committers to rubygems team. mame: Will reply. [Feature #16255] Make monitor.rb built-in (ko1) maybe there is no problem. Should we make MonitorMixin built-in? Discussion: matz: to make built-in expresses the wrong message to users to encourage to use threads ko1: ok, make it as extention library ko1: is Thread::Monitor necessary? usa: if it’s built-in, it’s necessary. but not, no. [Feature #16254] MRI internal: Define built-in classes in Ruby with intrinsic syntax (ko1) only on small start, maybe there is no problem. many points we need to decide. Discussion: TracePoint: https://gist.github.com/ko1/969e5690cda6180ed989eb79619ca612 Comparable: https://gist.github.com/ko1/7f18e66d1ae25bb30c7e823aa57f0d31 Need two files? Boot time performance. miniruby dependencies to external files. intrinsic doesn’t look great => keep using intrinsic for now. matz: no consideration because it’s an implementation detail but bigger than my expectation. akr: How about conditional compilation? We can use #ifdef freely now but Ruby file cannot refer such C-level constants. naruse: should release with 2.7. How will the ruby file path look like from ruby? => same as prelude internal:trace_point.rb. Timeline: will be merged within a few weeks, before preview 3. [Feature #16253] Shorthand “forward everything” syntax (mame) How about introducing def foo(…); bar(…); end? It does not solve the compatibility issue of keyword argument separation, but it would provide a useful shorthand for delegation if the code doesn’t have to work on 2.6. Discussion: You are allowed to use … as many times as you want in one method body. def parse_args(*args) args end def foo(…) args = parse_args(…) bar(…) End

block parameter cannot be … (implementation limitation) synchronize do |…| # SyntaxError end

OK: Valid only if … appears as a method parameter and it is parenthesized. Can be used multiple times: def foo(…); bar(…); baz(…); end p = proc {|a,b,c| }; p.call(…) def foo(…) super(…) end def foo(…); def bar(…); baz(…); end; end foo. p.(…) # p.call(…) def foo(…) iter{bar(…)}; end def foo(…) end # no error even if not used def foo(…); proc { bar(…) }; end; foo.call() NG: (at least in the initial implementation) Unavailable for block parameters. def foo(x, y, z) bar(…); end # Usable only if declared in the parameter list. def foo(x, y, z) super(…); end # ditto yield(…) (because yield does not accept a block) return (…) (ditto) Use as value a = … […] p = proc {|a,b,c| }; p[…] Array[…] foo[…] = x foo(…) { } (can’t call with a block) define_method(:foo) { |…| super(…) } iter{|…|} def foo(…): defined?(…); end # syntax error Without (): foo … (becomes Range) OK if implementable def foo(…); eval(‘bar(…)’); end Misc: No future for begin-less && end-less range. … should be meaningless; one could use .. instead. arity is -1 just like def f() parameters: FYI: p proc{}.parameters #=> [] FYI: def f() end; method(:f).parameters #=> [[:rest]] def f(…) end; method(:f).parameters #=> ? def foo(…) instance_eval(&Proc.new); end; def bar(…) foo { p(…) }; end; bar Future work: lead argument handling is postponed lead arguments can be extracted lead arguments can be added def f(x, y, …); g(1, 2, …); end [Feature #16150] Add a way to request a frozen string from to_s (mame) Symbol#to_s started to return a frozen string in 2.7.0-preview1, and that revealed at least seven incompatibility issues. To make sure: is it okay to keep this change in 2.7.0-preview2? Issues reported from some gems including middleman, tested with trunk. More issues will be reported after preview2. Discussion: mame: Fixing the issues is not very difficult. amatsuda: Rails fixed. naruse: negative. If unmaintained gems had a problem with this, what should one do? matz: Go ahead anyway. Get them fixed. [Feature #16120] Omitted block argument if block starts with dot-method call (Dan0042) last time Matz said “Give me time to consider it”; there is now a patch ready; is it ok to accept? Discussion: matz: No, I decided to reject it. Use the _n parameter. [Feature #13083] {String|Symbol}#match{?} with nil returns falsy as Regexp#match{?} (znz) Matz said “Those methods (but =~) should consistently raise exceptions.” and pull request’s conflicts resolved again. Discussion: matz: go ahead [Misc #16258] Combine call info and cache to speed up method invocation (alanwu) Looking for feedback and reviews. I think this offers a good perf boost. Discussion: ko1: will consider and reply — time up — [Bug #15912] Allow some reentrancy during TracePoint events (alanwu) I have seen quite a few people confused about Byebug’s REPL not working as they would expect due to this issue. I think it’s important to come up with some solution for this in two seven. Discussion: [Feature #16142] Implement code_range in Proc and Method - Ruby master - Ruby Issue Tracking System (osyo) Propose an API to get code position of Proc and Method so that we can get body of them (especially of a Proc). I want to discuss method names and return values Discussion: