To clarify, this isn't fixed because the merged PR https://github.com/ruby/ruby/pull/18209 was reverted in https://github.com/ruby/ruby/pull/18226. I'll close this as Rejected since there is no agreement how to fix this and at this po...Eregon (Benoit Daloze)
Applied in changeset commit:git|16ee033bece10f1d94b75a18d31f2d5c18c36d8a. ---------- [Feature #22520] Return the block node from #syntax_tree for a block Proc#syntax_tree, Method#syntax_tree and UnboundMethod#syntax_tree now return the...Eregon (Benoit Daloze)
Proc#syntax_tree, Method#syntax_tree and UnboundMethod#syntax_tree now return the BlockNode (or the SCOPE node with parse.y) for a proc or method defined by a block, instead of the call that the block is passed to. The call is not part o...Eregon (Benoit Daloze)
Such a proc is an ifunc block wrapping the method, so vm_proc_iseq() returned NULL and RubyVM::InstructionSequence.of returned nil for it, and so did Proc#syntax_tree, while Proc#source_range, Proc#source_location and Proc#parameters alr...Eregon (Benoit Daloze)
For an ArgumentError raised when calling a lambda or a method defined with define_method, the backtrace location of the callee is the block frame itself, so the node of that location is the node of the block iseq. Ruby 4.1 changes RubyV...Eregon (Benoit Daloze)
Thank you. I created https://github.com/ruby/prism/issues/4379 on the prism tracker to discuss how to get the parent/root node from a Prism node.Eregon (Benoit Daloze)
hsbt (Hiroshi SHIBATA) wrote: > The mswin platform, which is what `ruby/ruby` CI actually builds and tests with MSVC has had no end-user distribution at all. This is not quite accurate as there have been many efforts by @MSP-Greg (an...Eregon (Benoit Daloze)
Applied in changeset commit:git|db22451d7e1fcb04f033436fe099b0f0489d70d6. ---------- [Bug #22539] Do not apply refinements to the initialize call of opt_new opt_new inlines the call to initialize in the caller's bytecode, which made re...Eregon (Benoit Daloze)
opt_new inlines the call to initialize in the caller's bytecode, which made refinements of #initialize active at the call site of .new apply, unlike Class#new (e.g. via #send) and Ruby <= 3.4. Add the internal VM_CALL_IGNORE_REFINEMENTS...Eregon (Benoit Daloze)