@matz Could you approve these specific tweaks to the original proposal? * `Proc#source_range` starts from the `{`/`do` for `proc { foo }`/`proc do foo; end`, for the reasons mentioned in the previous comment * I added `Ruby::SourceRang...Eregon (Benoit Daloze)
PR is ready: https://github.com/ruby/ruby/pull/16835 When implementing I realized it's not a good idea to have the start of the `Ruby::SourceRange` to be the `p` of `proc { foo }`. It should be the `{`/`do` for a bunch of reasons: * Tha...Eregon (Benoit Daloze)
Concretely allowing Strings means extra checks for unknown or missing members, with only Symbols a `kwargs.size == CLASS_MEMBERS.size` check would be enough. It also leads to ambiguous cases like: ```ruby MyData = Data.define(:value) MyD...Eregon (Benoit Daloze)
This sounds like a serious eCryptfs bug to me, could you report it there? (or is there already an open issue about that? If so please link it here). (IMO eCryptfs is kind of a hack leaking many implementation details, nevertheless it'...Eregon (Benoit Daloze)
@nobu Was it intentional to keep accepting `String` and `to_str` kwargs in addition to Symbols? From what matz said above I would expect only Symbols are accepted, which would be valuable to optimize `Data#initialize` better.Eregon (Benoit Daloze)
shebang that includes "ruby" (https://github.com/ruby/prism/pull/4110) * Clarify the documentation of -x, it always searches a shebang that includes "ruby" * Update one more occurence --------- https://github.com/ruby/prism/commit/0c...Eregon (Benoit Daloze)
The implementation is still the same as of today: https://github.com/ruby/ruby/blob/4d87d43b01dbb312eb1ff5fbbc6c9f33218d91a2/thread.c#L2100-L2115 Meanwhile, `ruby_thread_has_gvl_p()` became a public API since #20877 ([commit](https:/...Eregon (Benoit Daloze)
Eregon (Benoit Daloze)08:46 PM Ruby Feature #21998: Add {Method,UnboundMethod,Proc}#source_range