Bug #9457: Regexp: when using subexpression calls, returned match groups seem wrong

• Status: Rejected
• Assignee: -
• Target version: -
• ruby -v: ruby 2.0.0p247 (2013-06-27 revision 41674) [x86_64-darwin12.4.0]
• Backport: 1.9.3: UNKNOWN, 2.0.0: UNKNOWN, 2.1: UNKNOWN

[dstosik (David Stosik)]

Hello,

Here are a two examples:

/(Tarzan|Jane) loves (\g<1>)/.match("Tarzan loves Jane").to_a
#=> ["Tarzan loves Jane", "Jane", "Jane"]
# Expected: ["Tarzan loves Jane", "Tarzan", "Jane"]

(http://rubular.com/r/We2x4XNB2l)

/(?<name>(Tarzan|Jane)) loves (\g<name>)/.match("Tarzan loves Jane").to_a
#=> ["Tarzan loves Jane", "Jane"]
# Expected: ["Tarzan loves Jane", "Tarzan", "Jane"]

(http://rubular.com/r/1e3oevc1Yg)

Unless I'm wrong on the results I'm expecting, I think this is a bug.

Thanks for your help,

David

[dstosik (David Stosik)]

"Subexpression calls":http://www.ruby-doc.org/core-2.0/Regexp.html#class-Regexp-label-Subexpression+Calls

[akr (Akira Tanaka)]

• Status changed from Open to Rejected

I think it is not a bug.

The subexpression, (Tarzan|Jane), matches twice: "Tarzan" and "Jane".
So the last matched string, "Jane", is recorded in the match object.

Consider following.

% ruby -e 'p /((Tarzan|Jane)) loves (\g<2>)/.match("Tarzan loves Jane")'
#<MatchData "Tarzan loves Jane" 1:"Tarzan" 2:"Jane" 3:"Jane">
% ruby -e 'p /(?<subj>(?<name>Tarzan|Jane)) loves (?<obj>\g<name>)/.match("Tarzan loves Jane")'
#<MatchData "Tarzan loves Jane" subj:"Tarzan" name:"Jane" obj:"Jane">