Jun 2017 Meeting

calendar_today June 16, 2017 confirmation_number 54 tickets discussed Source English

DevMeeting-2017-06-16

Date: 2017/06/16 (Fri) Time: 14:00- 18:00 (JST) Place: Money Forward inc. Headquarter Sign-up: https://ruby.connpass.com/event/57968/ Attendees: duerst (Martin Dürst) 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)

  • [Feature #12733] Bundle bundler to ruby core (shyouhei) updates?
  • [Feature #13332] Forwardable#def_instance_delegator nil (shyouhei)
  • [Feature #13333] block to yield (shyouhei)
  • [Feature #13378] Eliminate 4 of 8 syscalls when requiring file by absolute path (shyouhei)
  • [Feature #13381] [PATCH] Expose rb_fstring and its family to C extensions (shyouhei)
  • [Feature #13385] [PATCH] Make Resolv::DNS::Name validation similar to host and dig commands (shyouhei)
  • [Feature #13383] [PATCH] Module#source_location (shyouhei)
  • [Feature #12063] KeyError#receiver and KeyError#name (shyouhei) status?
  • [Bug #13397] #object_id should not be signed (shyouhei)
  • [Feature #13396] Net::HTTP has no write timeout (shyouhei)
  • [Bug #13407] We have recv_nonblock but not send_nonblock… can we add it? (shyouhei)
  • [Feature #13395] Add a method to check for not nil (shyouhei)
  • [Feature #13383] [PATCH] Module#source_location (shyouhei)
  • [Feature #13434] better method definition in C API (shyouhei)
  • [Feature #13512] System Threads (shyouhei)
  • [Feature #13518] Indented multiline comments (shyouhei)
  • [Feature #13516] Improve the text of the circular require warning (shyouhei)
  • [Feature #13532] Enable :encoding key or open-uri (open()) similar as to how File.read() and File.readlines() already allow for (shyouhei)
  • [Bug #13535] Installing Ruby2.4.1 on Solaris 10 (shyouhei) response?
  • [Feature #13568] File#path for O_TMPFILE fds are unmeaning (sorah)
  • [Feature #13563] Implement Hash#choice method. (shyouhei)
  • Previous bugs that were not assigned (shyouhei)
    • [Bug #13196] Improve keyword argument errors when non-keyword arguments given
    • [Bug #13320] rescue blocks get an entry in backtrace locations
    • [Bug #13336] Default Parameters don’t work
    • [Bug #13337] Eval and Later Defined Local Variables
    • [Bug #13350] File.read :newline option not respected on Linux
    • [Bug #13373] FileUtils methods for copy, move and remove directories is not providing a decent error trace for letting know if it was success or fail
    • [Bug #13101] Date#rfc2822 and Time#rfc2822 don’t return the same format
    • [Bug #12684] Delegator#eql? missing
    • [Feature #13389] [PATCH] POP3 support timeout for TLS handshake
    • [Bug #13404] Hash#any? yields arguments to lambdas with proc semantics
    • [Bug #13413] –with-static-linked-ext doesn’t install extension files on make install
    • [Bug #13429] Net::SMTP has no read timeout when connexion over TLS
    • [Bug #13432] set_trace_funcにproc->is_from_method = TRUEのオブジェクトを渡し、SystemStackErrorを発生させるとRubyVMが停止する
    • [Bug #13446] refinements with prepend for module has strange behavior
    • [Bug #13498] Weakref, Weakmap and define_finalizer don’t work on frozen objects
    • [Bug #13513] Resolv::DNS::Message.decode hangs after detecting truncation in UDP messages
    • [Bug #13501] Process.kill behaviour for negative pid is not documented and may be wrong
    • [Bug #13515] Pathname#join doesn’t add separator on UNC paths
    • [Bug #13521] [PATCH] Add fallback for DNS resolver registry key on Wine
    • [Bug #13537] ruby crash in rb_gc_mark
    • [Bug #13548] miniruby SEGV while building with non-default CFLAGS (caused by __builtin_setjmp)
    • [Bug #13555] Disable TestTrace#test_trace_stackoverflow
    • [Bug #13557] there’s no way to pass backtrace locations as a massaged backtrace
    • [Bug #10290] segfault when calling a lambda recursively after rescuing SystemStackError
      • Fix confirmed?
    • [Feature #10674] Net::HTTP retries idempotent requests once after a timeout, but its not configurable
    • [Bug #13542] MinGW trunk Builds - Summary of Issues
      • [Bug #13549] MinGW / Windows encoding - Two issues
      • [Bug #13556] MinGW readline Alt / Meta keys
      • [Bug #13569] Windows - TestRubyOptions#test_search - append to paths instead of replacing
    • [Bug #13564] Exception message management

From attendees

  • [Bug #12159] Thread::Backtrace::Location#path returns absolute path for files loaded by require_relative (ko1)
  • [Bug #13576] File#to_path shall be deleted (shyouhei)

From non-attendees

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

  • [Feature #13618] [PATCH] auto fiber schedule for rb_wait_for_single_fd and rb_waitpid (normalperson, ko1)
  • [Feature #12694] Want a String method to remove heading substr.
    • (sonots) I implemented with a name String#remove_prefix. If the name is ok, I want to merge it. Matz, could you decide? There were other candidates such as lchomp, trim_prefix. I dropped trim_prefix because trim is used to remove a list of characters rather than a substring in some languages, so I felt trim_prefix would raise confusion. I dropped lchomp because I just had never heard.

Log

Date: 2017/06/16 (Fri)

Time: 14:00- 18:00 (JST) Place: Money Forward inc. Headquarter Sign-up: https://ruby.connpass.com/event/57968/ log edit: https://docs.google.com/document/d/1z19pKt8jlpiEUR3RnWWBCfs3OR_hbiAZMwpQ6ZTllP0/edit# Attendees: duerst (Martin Dürst) Language: mostly Japanese (sorry for non native Japanese speakers)

Agenda

  • next

  • 7/14, 14:00~

  • at Cookpad HQ, Yebisu

  • next after

  • 8/31
  • at Cookpad HQ, Yebisu

About 2.5 timeframe

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

Carry-over from previous meeting(s)

[Feature #12733] Bundle bundler to ruby core (shyouhei) updates?

  • hsbt: rubygems/rubygems 2.7 requires bundler.  After it is released, when ruby-core merges that version, we have to take care.

  • import bundler for test then?

  • if so why only for test?

  • ko1: is there any blocker?
  • hsbt: rspec.

[Feature #13332] Forwardable#def_instance_delegator nil (shyouhei)

  • assign nobu

[Feature #13333] block to yield (shyouhei)

  • ko1: we need more use cases.
  • martin: nobu: add them to the ticket.

[Feature #13378] Eliminate 4 of 8 syscalls when requiring file by absolute path (shyouhei)

  • nobu: I don’t want to handle fds this way because we can’t close properly on exceptions. I think we should wrap them using IOs.

[Feature #13381] [PATCH] Expose rb_fstring and its family to C extensions (shyouhei)

  • ko1: how about it? I don’t like “f”string naming.
  • nobu: other languages call it “intern”
  • ko1: true but we already have rb_intern().
  • ko1: nobody has idea. so that nobu and ko1 will decide it. nobody against exposing this functionality.

[Feature #13385] [PATCH] Make Resolv::DNS::Name validation similar to host and dig commands (shyouhei)

  • akr: I can’t say if this is valid until I fully understand the RFC.
  • akira: what about using domain_name gem?
  • assign akr.

[Feature #13383] [PATCH] Module#source_location (shyouhei)

  • akira: needs?
  • shyouhei: I think source_locations makes more sense than source_location.
  • akira: I started to confuse.  Is it useful?  We can already get Method#source_location.

[Feature #12063] KeyError#receiver and KeyError#name (shyouhei) status?

  • matz: I take KeyError#key.

[Bug #13397] #object_id should not be signed (shyouhei)

  • akr: isn’t it possible to somehow calculate so that we can avoid negative object_id?  Because pointers are 8 byte aligned, while positive fixnums are of 262 space.
  • nobu: it’s true we can avoid negative numbers on some environment, but not always.
  • reject this specific issue and request for new issue that address the inspect helper.

[Feature #13396] Net::HTTP has no write timeout (shyouhei)

[Bug #13407] We have recv_nonblock but not send_nonblock… can we add it? (shyouhei)

[Feature #13395] Add a method to check for not nil (shyouhei)

[Feature #13434] better method definition in C API (shyouhei)

  • nobu: I think parsing string is not a good idea.

[Feature #13512] System Threads (shyouhei)

[Feature #13518] Indented multiline comments (shyouhei)

[Feature #13516] Improve the text of the circular require warning (shyouhei)

[Feature #13532] Enable :encoding key or open-uri (open()) similar as to how File.read() and File.readlines() already allow for (shyouhei)

[Bug #13535] Installing Ruby2.4.1 on Solaris 10 (shyouhei) response?

[Feature #13568] File#path for O_TMPFILE fds are unmeaning (sorah)

[Feature #13563] Implement Hash#choice method. (shyouhei)

  • Previous bugs that were not assigned (shyouhei)

[Bug #13196] Improve keyword argument errors when non-keyword arguments given

[Bug #13320] rescue blocks get an entry in backtrace locations

[Bug #13336] Default Parameters don’t work

[Bug #13337] Eval and Later Defined Local Variables

[Bug #13350] File.read :newline option not respected on Linux

[Bug #13373] FileUtils methods for copy, move and remove directories is not providing a decent error trace for letting know if it was success or fail

[Bug #13101] Date#rfc2822 and Time#rfc2822 don’t return the same format

[Bug #12684] Delegator#eql? missing

[Feature #13389] [PATCH] POP3 support timeout for TLS handshake

[Bug #13404] Hash#any? yields arguments to lambdas with proc semantics

[Bug #13413] –with-static-linked-ext doesn’t install extension files on make install

[Bug #13429] Net::SMTP has no read timeout when connexion over TLS

[Bug #13432] set_trace_funcにproc->is_from_method = TRUEのオブジェクトを渡し、SystemStackErrorを発生させるとRubyVMが停止する

[Bug #13446] refinements with prepend for module has strange behavior

[Bug #13498] Weakref, Weakmap and define_finalizer don’t work on frozen objects

[Bug #13513] Resolv::DNS::Message.decode hangs after detecting truncation in UDP messages

[Bug #13501] Process.kill behaviour for negative pid is not documented and may be wrong

[Bug #13515] Pathname#join doesn’t add separator on UNC paths

[Bug #13521] [PATCH] Add fallback for DNS resolver registry key on Wine

[Bug #13537] ruby crash in rb_gc_mark

[Bug #13548] miniruby SEGV while building with non-default CFLAGS (caused by __builtin_setjmp)

[Bug #13555] Disable TestTrace#test_trace_stackoverflow

[Bug #13557] there’s no way to pass backtrace locations as a massaged backtrace

[Bug #10290] segfault when calling a lambda recursively after rescuing SystemStackError

  • Fix confirmed?

  • nobu: the variant of setjmp seems to be the root cause.  The proposed workaround works, but may impact performance.  We have to measure before changing the default.

[Feature #10674] Net::HTTP retries idempotent requests once after a timeout, but its not configurable

[Bug #13542] MinGW trunk Builds - Summary of Issues

[Bug #13549] MinGW / Windows encoding - Two issues

[Bug #13556] MinGW readline Alt / Meta keys

[Bug #13569] Windows - TestRubyOptions#test_search - append to paths instead of replacing

[Bug #13564] Exception message management

From attendees

[Bug #12159] Thread::Backtrace::Location#path returns absolute path for files loaded by require_relative (ko1)

  • ko1: I want to add Thread::Backtrace::Location#realpath to deprecate #absplute_path instead.
  • matz: no objection.

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

  • [Bug #13576] File#to_path shall be deleted (shyouhei)

  • akr: did you try deleting this method?
  • shyouhei: no, but I should.
  • akr: we must start from adding warnings.

From non-attendees

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

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

  • ko1: programmers transfer execution of fibers by hand.  Eric proposes to introduce “auto”-Fiber which are OK to transfer each other on blocking situations. His implementation is great. rb_wait_for_single_fd is the ponit of switching. matz wanted this feature, too.
  • shyouhei: can auto Fibers be a subject to transfer or an object?
  • ko1: transfer happens only from an auto fiber to another and nothing else, AFAIK.
  • ko1: This is essentially a re-introduction of Ruby 1.8 threads (without time-tick). I’m concerning about thread difficulties like locking.
  • akr: is locking OK for auto Fibers?
  • ko1: Fibers has not needed locking because so far, Fiber.yield has been the only mechanism to transfer execution so it was very explicit.  auto-Fiber changes that so there can be unintended transfers.
  • matz: do you want to reject this?
  • ko1: I don’t want to accept until we introduce a way to explicitly state where we can transfer, like async/await.
  • ko1: I think it’s NG to call routines that are not Fiber-aware from proposed  auto-Fiber.  I don’t think people can insert appropriate Thread.exclusive calls because if they could, we could never have experienced resource contentions.
  • matz: I think it’s hard to tag every and all of fiber-aware routines to be async.  That sounds too big.
  • matz: show us your counter-proposal.
  • ko1: is it OK for you to accept Eric’s proposal as-is?
  • matz: I’m positive, but I know your concern so I won’t accept right now.

[Feature #12694] Want a String method to remove heading substr.

  • (sonots) I implemented with a name String#remove_prefix. If the name is ok, I want to merge it. Matz, could you decide? There were other candidates such as lchomp, trim_prefix. I dropped trim_prefix because trim is used to remove a list of characters rather than a substring in some languages, so I felt trim_prefix would raise confusion. I dropped lchomp because I just had never heard.

  • candidates discussed:

  • lchomp
  • chompl
  • guilotine

  • ltrim / rtrim

  • NG because PHP behaves differently

  • delete_prefix

  • amatsuda: remove_prefix sounds strange if it doesn’t change the receiver.
  • nobu: facets has lchomp.
  • akr: I don’t want chomp’s complex newline handling for this requested method.
  • ko1: remove_prefix seems has no downsides.
  • matz: accept delete_prefix.