Misc #18557: DevMeeting-2022-02-17

• Status: Closed
• Assignee: -

[mame (Yusuke Endoh)]

The next dev meeting

Date: 2022/02/17 13:00-17:00 (JST)
Log: https://github.com/ruby/dev-meeting-log/blob/master/DevMeeting-2022-02-17.md

• 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 record of the discussion in the 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 of the meeting are scheduled according to when/where we can reserve Matz's time.
• DO NOT discuss then on this ticket, please.

Call for agenda items

If you have a ticket that you want matz and committers to discuss, please post it into this ticket in the following format:

* [Ticket ref] Ticket title (your name)
  * Comment (A summary of the ticket, why you put this ticket here, what point should be discussed, etc.)

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.

• It is recommended to add a comment by 2022/02/14. We hold a preparatory meeting to create an agenda a few days before the dev-meeting.
• The format is strict. We'll use this script to automatically create an markdown-style agenda. We may ignore a comment that does not follow the format.
• Your comment is mandatory. We cannot read all discussion of the ticket in a limited time. We appreciate it if you could write a short summary and update from a previous discussion.

[mame (Yusuke Endoh)]

• Related to Misc #14770: [META] DevelopersMeeting added

[shugo (Shugo Maeda)]

• [Feature #13110] Byte-based operations for String
  • Are there any objections to introduce byteindex, byterindex and byteoffset? (not bytesplice)
  • Should the method names be byte_index, byte_rindex, and byte_offset?

[peterzhu2118 (Peter Zhu)]

• [Feature #18249] The ABI version of dev builds of CRuby does not correspond to the ABI
  • ABI checking for development Ruby.
  • Calculates the MD5 hash of header files during compilation of Ruby, stores this MD5 in native extensions, and compares the values when loading the binary.
  • Prevents incompatible binaries from being loaded, which can cause bugs, crashes, or other incorrect behavior.
  • Summary of motivation here: https://bugs.ruby-lang.org/issues/18249#note-15

[hsbt (Hiroshi SHIBATA)]

• [Feature #18513] Hide patchlevel from ruby -v
  • There is no reason to display it now.

[byroot (Jean Boussier)]

• [Feature #18498] Introduce a public WeakKeysMap that compares by equality (byroot)
  • Ruby's current weak reference constructs are not very usable.
  • The most useful weak collection that is offered by most platform is WeakKeysMap
  • The keys would be weak references, but the values would be strong references
  • It would use equality semantic, just like a regular Hash

[headius (Charles Nutter)]

• [Feature #18568] Explore lazy RubyGems boot to reduce need for --disable-gems
  • With more and more default gems, RubyGems has become a major part of Ruby's startup time.
  • On CRuby, the base boot time is reduced by 80% using --disable-gems.
  • Several ruby-core members would like to remove --disable-gems but others and some users need it for fast command line tools.
  • It seems like it is time to eliminate the overhead of booting RubyGems and default gemspecs.

[hsbt (Hiroshi SHIBATA)]

• [Feature #18571] Removed the bundled sources from release package after Ruby 3.2
  • (dummy comment)

[byroot (Jean Boussier)]

[Feature #18566] Merge io-wait gem into core IO class. (byroot)

• It's very small and any non-trivial IO code will require it.
• Merge io-wait into io.c for Ruby 3.2
• Remove io-wait as a dependency of all gems maintained by ruby-core (e.g. net-protocol).
• Publish a new io-wait version that is simply an empty gem.
• Add a lib/io/wait.rb stub, with eventually a deprecation warning.
• @normalperson (Eric Wong) suggest to do the same with io-nonblock.

[normalperson (Eric Wong)]

"byroot (Jean Boussier)" [e-mail] wrote:

Issue #18557 has been updated by byroot (Jean Boussier).

[Feature #18566] Merge io-wait gem into core IO class. (byroot)

• It's very small and any non-trivial IO code will require it.
• Merge io-wait into io.c for Ruby 3.2
• Remove io-wait as a dependency of all gems maintained by ruby-core (e.g. net-protocol).
• Publish a new io-wait version that is simply an empty gem.
• Add a lib/io/wait.rb stub, with eventually a deprecation warning.

Please just keep lib/io/wait.rb stub empty indefinitely. No
need for deprecation warning, just a no-op is fine and don't
cause pain for existing users. I don't want to hurt:

1. users who aren't coders
2. users who lack permissions to change code
3. codebases which need to support both old/ancient Rubies + new ones

I'm looking at system-wide implications from distro-installed Rubies;
not from a webapp+bundler perspective.

• @normalperson (Eric Wong) suggest to do the same with io-nonblock.

Actually, Benoit suggested io-nonblock. I suggested going beyond that :)

[jeremyevans0 (Jeremy Evans)]

• [Bug #18487] Kernel#binding behaves differently depending on implementation language of items on the stack (jeremyevans0)
  • Is it OK to limit Kernel#binding to only look at the preceding frame, instead of the closest Ruby-level frame?
  • If so, is the patch acceptable?
• [Bug #16663] Add block or filtered forms of Kernel#caller to allow early bail-out (jeremyevans0)
  • Since the last developer meeting, a prototype has been created.
  • Is the behavior of the prototype acceptable?
  • Incorrect usage with to_enum raises StopIteration instead of crashing. Is that acceptable?
• [Bug #14103] Regexp absense operator has no chance to ^C (jeremyevans0)
  • One possible way around this is checking for interrupts and potentially yield to another thread every N times during backtracking.
  • I submitted a pull request to do that (N = 1000), and it fixes the example given.
  • Is such an approach acceptable? If so, is the patch acceptable?
  • One issue not addressed by the patch is the modification of the string being searched by another thread during regexp evaluation (potentially causing a crash), but I'm not sure whether that is possible.

[greggzst (Grzegorz Jakubiak)]

• [Feature #16989] Sets: need ♥️
  • matz has already accepted adding Sets as default. However, there hasn’t been any updates on the progress

[byroot (Jean Boussier)]

• [Feature #18339] GVL instrumentation API
  • The goal is to be able to measure GVL contention and tell users if they use too many threads for their workload.
  • It would allow to instrument the GVL which is a key metric for threaded environments, and to tune concurrency in applications.
  • I implemented a (posix only for now) draft: https://github.com/ruby/ruby/pull/5500.
  • The overhead when no hook is registered is just a single unprotected boolean check, so close to zero.
  • rb_gvl_hook_t * rb_gvl_event_new(rb_gvl_callback callback, rb_event_flag_t event) to register a hook.
  • bool rb_gvl_event_delete(rb_gvl_hook_t * hook) to unregister it.
  • Is the general idea acceptable?
  • Any API change before I try to implement the Windows version?

[tenderlovemaking (Aaron Patterson)]

• [Feature #12962] Adding "friend" method visibility
  • Allow libraries with Object Oriented internals to call "not public" methods
  • Useful when "protected" is too restricted, but the method shouldn't be "public"
  • Refinements can't help because they are lexically scoped
    • It's easier to add method visibility than to move all code around

[zverok (Victor Shepelev)]

• [Bug #18580] Range#include? inconsistency for beginless String ranges (zverok)
  • (...'z').include?('ww') works for strings, due to historical reasons
  • there is no semantical justification for it, and it is confusing for users

[ktsj (Kazuki Tsujimoto)]

• [Feature #18585] Promote find pattern to official feature (ktsj)
  • Is it OK to to promote find pattern to official feature?

[nobu (Nobuyoshi Nakada)]

• [Misc #18362] mswin builds & vs2022
  • What can we do for the bug?

[byroot (Jean Boussier)]

• [Feature #18576] Rename ASCII-8BIT encoding to BINARY
  • The ASCII-8BIT name is very confusing for beginners.
  • Even for people who know what it is, it can easily be misread as US-ASCII
  • It has been aliased as BINARY for a very long time, and in the vast majority of the case it's a better description.
  • I expect the backward compatibility problems to be very minimal (limited to Encoding#name).

[kddnewton (Kevin Newton)]

• [Feature #18589] Finer-grained constant invalidation (Kevin Newton)
  • We want to trade a very small memory increase for much better constant cache invalidation.
  • This will allow relying on more stable constant caches for better code specialization in JITs.
  • This can also help speed up initial boot times since caches are currently invalidated each time a class is created.

[mame (Yusuke Endoh)]

• Description updated (diff)
• Status changed from Open to Closed