--- !ditz.rubyforge.org,2008-03-06/issue 
title: "regexp non-capturing group bug: /(?:a)?/.exec('x') does not match (but should)"
desc: /a?/ matches, so does /(a)?/
type: :bugfix
component: duk
release: v0.10
reporter: sva <sami.vaarala@iki.fi>
status: :closed
disposition: :fixed
creation_time: 2014-02-19 18:32:29.242684 Z
references: []

id: c259c7a979801e1ae101231d116ea8c77f40a98c
log_events: 
- - 2014-02-19 18:32:29.617574 Z
  - sva <sami.vaarala@iki.fi>
  - created
  - ""
- - 2014-02-19 18:34:17.036317 Z
  - sva <sami.vaarala@iki.fi>
  - commented
  - |-
    Regexp bytecode for /a?/
    
    /a?/
    \x00    flags
    \x02    nsaved
    \x0b    save
    \x00    index
    \x0a    sqgreedy
    \x00    min
    \x01    max
    \x01    atomlen
    \x06    skip = +3
    \x02    char
    a       'a'
    \x01    match
    \x0b    save
    \x01    match
    \x01    match
    
    Regexp bytecode for /(?:a)?/ when compiled with the bug present:
    
    \x00    flags
    \x02    nsaved
    \x0b    save
    \x00    index
    \x02    char
    a       'a'
    \x0a    sqgreedy
    \x00    min
    \x01    max
    \x01    atomlen
    \x02    skip = +1
    \x01    match
    \x0b    save
    \x01    index
    \x01    match
    
    The bug in the regexp compiler causes the regexp to be essentially
    compiled into /a(?:)?/.
    
    After a potential fix, the bytecode is identical to /a?/:
    
    \x00    flags
    \x02    nsaved
    \x0b    save
    \x00    0
    \x0a    sqgreedy
    \x00    min
    \x01    max
    \x01    atomlen
    \x06    skip = +3
    \x02    char
    a       'a'
    \x01    match
    \x0b    save
    \x01    index
    \x01    match
- - 2014-02-19 22:35:54.814052 Z
  - sva <sami.vaarala@iki.fi>
  - closed with disposition fixed
  - ""