Modify ↓

Opened 11 years ago

Last modified 8 months ago

#2696 assigned Bug

StringRegExp* return non-participating groups

Reported by: jchd18 Owned by: Jon
Milestone: Component: AutoIt
Version: 3.3.11.4 Severity: None
Keywords: Cc:

Description

Non-participating groups appear as part of the result of our PCRE wrappers, but they should ignore them.

Long version: pattern subroutines created by (?(DEFINE) ...) are named groups for PCRE, thus get numbered as usual and do appear in the ovector parameter of pcre_exec() internal function.

Of course they never actually hold any result since the corresponding ovector tuple is (-1, -1). This is different from when actual capturing groups match an empty string: their ovector tuple has length = 0. We sould test for (-1, -1) and ignore these meaningless entries.

See ​http://www.autoitscript.com/forum/topic/160802-stringsplit-multiple-whole-words-autoit/#entry1167381 for an example.

PM jchd if needed.

Attachments (0)

Change History (9)

comment:1 Changed 11 years ago by jchd18

BTW the same effect occurs with groups having repetition count {0} for instance:
Pattern:

(?x) (a){0} (a?) (bc) (?(DEFINE) (?<head> x)) (?(?=$) z?)

Subject:

bc
abc

In this case RegexBuddy clearly shows the non-participating groups when using full detail mode:
Result:

                             Start   Length
Match 1:	bc	     0	     2
Group 1 did not participate in the match
Group 2:		     0	     0
Group 3:	bc	     0	     2
Group "head" did not participate in the match
Match 2:	abc	     3	     3
Group 1 did not participate in the match
Group 2:	a	     3	     1
Group 3:	bc	     4	     2
Group "head" did not participate in the match

comment:2 Changed 11 years ago by jchd18

As a clue to help locate the failling logic, here's the demonstration that non-participating groups placed after all capturing groups don't cause the ghost capture bug. Only those appearing before or in between capturing groups trigger the issue.

#include <Array.au3>

Local $patterns = [ _
	["(?x)                          (\b\w+)", "No bug: simple capturing group"], _
	["(?x) (?(DEFINE)(?<head> xxx)) (\b\w+)", "Bug: non-participating group before capturing group"], _
	["(?x) (?(DEFINE)(?<head> xxx)) (\b\w+) (?(DEFINE)(?<tail> yyy))", "Bug: adding a non-participating group after the capturing group doesn't matter"], _
	["(?x) (?(DEFINE)(?<head> xxx))  \b\w+  (?(DEFINE)(?<tail> yyy))", "No bug: no capturing group"], _
	["(?x)                          (\b\w+) (?(DEFINE)(?<tail> yyy))", "No bug: non-participating group last"], _
	["(?x) (x){0} (y){0} (z){0}     (\b\w+)", "Bug: more non-participating groups first"] _
]
For $i = 0 To UBound($patterns) - 1
	$res = StringRegExp("There is a bug somewhere.", $patterns[$i][0], 3)
	_ArrayAdd($res, $patterns[$i][1])
	_ArrayDisplay($res, $patterns[$i][1])
Next

comment:3 Changed 4 years ago by Jpm

  • Owner set to Jpm
  • Status changed from new to assigned

Thanks jchd for testing the complete fix
fix sent to Jon

comment:4 Changed 3 years ago by Jon

  • Milestone set to 3.3.15.5
  • Owner changed from Jpm to Jon
  • Resolution set to Fixed
  • Status changed from assigned to closed

Fixed by revision [12632] in version: 3.3.15.5

comment:5 Changed 3 years ago by Jos

  • Resolution Fixed deleted
  • Status changed from closed to reopened

Please check ​https://www.autoitscript.com/forum/topic/207664-autoit-v33155-beta-issue-with-_pathsplit/?tab=comments#comment-1497690

Version 0, edited 3 years ago by Jos (next)

comment:6 Changed 3 years ago by TicketCleanup

  • Milestone 3.3.15.5 deleted

Automatic ticket cleanup.

comment:7 Changed 3 years ago by Jon

Change reverted.

comment:8 Changed 3 years ago by Jpm

  • Owner changed from Jon to Jpm
  • Status changed from reopened to assigned

I resent a fix to Jon,
I hope he wiil integrate it

comment:9 Changed 8 months ago by Jpm

  • Owner changed from Jpm to Jon

Guidelines for posting comments:

  • You cannot re-open a ticket but you may still leave a comment if you have additional information to add.
  • In-depth discussions should take place on the ​forum.

For more information see the full version of the ticket guidelines here.

Add Comment

Modify Ticket

Action
as assigned The owner will remain Jon.
Author


E-mail address and user name can be saved in the Preferences.

Attachments ↑ Description ↑
 
Note: See TracTickets for help on using tickets.