Bug #15640: Inconsistent value passed to respond_to_missing?

• Status: Closed
• Assignee: -
• Target version: -
• ruby -v: ruby 2.5.1p57 (2018-03-29 revision 63029) [x86_64-linux-gnu]
• Backport: 2.4: UNKNOWN, 2.5: UNKNOWN, 2.6: UNKNOWN

[kirun (Kieran Leigh)]

If the symbol has been defined, then it is passed. Otherwise a string is passed the first time.

class SomeBug
  def respond_to_missing?(m, *)
    p m
    true
  end
end
​
if false then
  # uncomment to force to symbol--location doesn't matter, just the literal parse:
  # :aa
end
​
sd = SomeBug.new
​
# NOTE: single character strings do not show this behavior
sd.method("aa".to_sym)       # "aa" unless :aa uncommented above
sd.method("aa".to_sym)       # :aa
sd.method("bb".to_sym)       # "bb"
sd.method("bb".to_sym)       # :bb

[nobu (Nobuyoshi Nakada)]

• Description updated (diff)

Good catch.
This behavior was to get rid of inadvertent symbol creation before symbol-GC,
so it is no longer needed.

[Hanmac (Hans Mackowiak)]

@nobu (Nobuyoshi Nakada) should the value depending on the value used for method, or should respond_to_missing? always has Symbol as parameter?

i think the best would be to make it simple and always have it use Symbol

[nobu (Nobuyoshi Nakada)]

• Status changed from Open to Closed

Applied in changeset trunk|r67185.

---

proc.c: call respond_to_missing? with a symbol

[ruby-core:91683] [Bug #15640]