# From laura_fairhead@talk21.com Fri May 10 11:24:41 2002
# Return-Path: <laura_fairhead@talk21.com>
# Received: from localhost (aahz [127.0.0.1])
# by skeeve.com (8.11.2/8.11.2) with ESMTP id g4A8OdU01822
# for <arnold@localhost>; Fri, 10 May 2002 11:24:40 +0300
# Received: from actcom.co.il [192.114.47.1]
# by localhost with POP3 (fetchmail-5.7.4)
# for arnold@localhost (single-drop); Fri, 10 May 2002 11:24:40 +0300 (IDT)
# Received: by actcom.co.il (mbox arobbins)
# (with Cubic Circle's cucipop (v1.31 1998/05/13) Fri May 10 11:30:42 2002)
# X-From_: laura_fairhead@talk21.com Fri May 10 05:39:57 2002
# Received: from lmail.actcom.co.il by actcom.co.il with ESMTP
# (8.11.6/actcom-0.2) id g4A2dpw26380 for <arobbins@actcom.co.il>;
# Fri, 10 May 2002 05:39:52 +0300 (EET DST)
# (rfc931-sender: mail.actcom.co.il [192.114.47.13])
# Received: from f7.net (consort.superb.net [209.61.216.22])
# by lmail.actcom.co.il (8.11.6/8.11.6) with ESMTP id g4A2dxl10851
# for <arobbins@actcom.co.il>; Fri, 10 May 2002 05:39:59 +0300
# Received: from fencepost.gnu.org (fencepost.gnu.org [199.232.76.164])
# by f7.net (8.11.6/8.11.6) with ESMTP id g4A2dwN11097
# for <arnold@skeeve.com>; Thu, 9 May 2002 22:39:58 -0400
# Received: from [194.73.242.6] (helo=wmpmta04-app.mail-store.com)
# by fencepost.gnu.org with smtp (Exim 3.34 #1 (Debian))
# id 1760K4-0001QX-00
# for <bug-gawk@gnu.org>; Thu, 09 May 2002 22:39:56 -0400
# Received: from wmpmtavirtual ([10.216.84.15])
# by wmpmta04-app.mail-store.com
# (InterMail vM.5.01.02.00 201-253-122-103-101-20001108) with SMTP
# id <20020510023921.EEW24107.wmpmta04-app.mail-store.com@wmpmtavirtual>
# for <bug-gawk@gnu.org>; Fri, 10 May 2002 03:39:21 +0100
# Received: from 213.1.102.243 by t21web05-lrs ([10.216.84.15]); Fri, 10 May 02 03:38:42 GMT+01:00
# X-Mailer: talk21 v1.24 - http://talk21.btopenworld.com
# From: laura_fairhead@talk21.com
# To: bug-gawk@gnu.org
# X-Talk21Ref: none
# Date: Fri, 10 May 2002 03:38:42 GMT+01:00
# Subject: bug in gawk 3.1.0 regex code
# Mime-Version: 1.0
# Content-type: multipart/mixed; boundary="--GgOuLpDpIyE--1020998322088--"
# Message-Id: <20020510023921.EEW24107.wmpmta04-app.mail-store.com@wmpmtavirtual>
# X-SpamBouncer: 1.4 (10/07/01)
# X-SBClass: OK
# Status: RO
#
# Multipart Message Boundary - attachment/bodypart follows:
#
#
# ----GgOuLpDpIyE--1020998322088--
# Content-Type: text/plain
# Content-Transfer-Encoding: 7bit
#
#
# I believe I've just found a bug in gawk3.1.0 implementation of
# extended regular expressions. It seems to be down to the alternation
# operator; when using an end anchor '$' as a subexpression in an
# alternation and the entire matched RE is a nul-string it fails
# to match the end of string, for example;
#
# gsub(/$|2/,"x")
# print
#
# input = 12345
# expected output = 1x345x
# actual output = 1x345
#
# The start anchor '^' always works as expected;
#
# gsub(/^|2/,"x")
# print
#
# input = 12345
# expected output = x1x345
# actual output = x1x345
#
# This was with POSIX compliance enabled althought that doesn't
# effect the result.
#
# I checked on gawk3.0.6 and got exactly the same results however
# gawk2.15.6 gives the expected results.
#
# All the follow platforms produced the same results;
#
# gawk3.0.6 / Win98 / i386
# gawk3.1.0 / Win98 / i386
# gawk3.0.5 / Linux2.2.16 / i386
#
# Complete test results were as follows;
#
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# regex input expected actual bug?
# -------------------------------------------------------------
# (^) 12345 x12345 x12345
# ($) 12345 12345x 12345x
# (^)|($) 12345 x12345x x12345x
# ($)|(^) 12345 x12345x x12345x
# 2 12345 1x345 1x345
# (^)|2 12345 x1x345 x1x345
# 2|(^) 12345 x1x345 x1x345
# ($)|2 12345 1x345x 1x345 **BUG**
# 2|($) 12345 1x345x 1x345 **BUG**
# (2)|(^) 12345 x1x345 x1x345
# (^)|(2) 12345 x1x345 x1x345
# (2)|($) 12345 1x345x 1x345 **BUG**
# ($)|(2) 12345 1x345x 1x345 **BUG**
# ((2)|(^)). 12345 xx45 xx45
# ((^)|(2)). 12345 xx45 xx45
# .((2)|($)) 12345 x34x x34x
# .(($)|(2)) 12345 x34x x34x
# (^)|6 12345 x12345 x12345
# 6|(^) 12345 x12345 x12345
# ($)|6 12345 12345x 12345x
# 6|($) 12345 12345x 12345x
# 2|6|(^) 12345 x1x345 x1x345
# 2|(^)|6 12345 x1x345 x1x345
# 6|2|(^) 12345 x1x345 x1x345
# 6|(^)|2 12345 x1x345 x1x345
# (^)|6|2 12345 x1x345 x1x345
# (^)|2|6 12345 x1x345 x1x345
# 2|6|($) 12345 1x345x 1x345 **BUG**
# 2|($)|6 12345 1x345x 1x345 **BUG**
# 6|2|($) 12345 1x345x 1x345 **BUG**
# 6|($)|2 12345 1x345x 1x345 **BUG**
# ($)|6|2 12345 1x345x 1x345 **BUG**
# ($)|2|6 12345 1x345x 1x345 **BUG**
# 2|4|(^) 12345 x1x3x5 x1x3x5
# 2|(^)|4 12345 x1x3x5 x1x3x5
# 4|2|(^) 12345 x1x3x5 x1x3x5
# 4|(^)|2 12345 x1x3x5 x1x3x5
# (^)|4|2 12345 x1x3x5 x1x3x5
# (^)|2|4 12345 x1x3x5 x1x3x5
# 2|4|($) 12345 1x3x5x 1x3x5 **BUG**
# 2|($)|4 12345 1x3x5x 1x3x5 **BUG**
# 4|2|($) 12345 1x3x5x 1x3x5 **BUG**
# 4|($)|2 12345 1x3x5x 1x3x5 **BUG**
# ($)|4|2 12345 1x3x5x 1x3x5 **BUG**
# ($)|2|4 12345 1x3x5x 1x3x5 **BUG**
# x{0}((2)|(^)) 12345 x1x345 x1x345
# x{0}((^)|(2)) 12345 x1x345 x1x345
# x{0}((2)|($)) 12345 1x345x 1x345 **BUG**
# x{0}(($)|(2)) 12345 1x345x 1x345 **BUG**
# x*((2)|(^)) 12345 x1x345 x1x345
# x*((^)|(2)) 12345 x1x345 x1x345
# x*((2)|($)) 12345 1x345x 1x345 **BUG**
# x*(($)|(2)) 12345 1x345x 1x345 **BUG**
# x{0}^ 12345 x12345 x12345
# x{0}$ 12345 12345x 12345x
# (x{0}^)|2 12345 x1x345 x1x345
# (x{0}$)|2 12345 1x345x 1x345 **BUG**
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#
#
# Here's the test program I used, a few of the cases use ERE {n[,[m]]}
# operators so need '-W posix', (although the same results minus
# those tests came out without POSIX compliance enabled)
#
# [ Invocation was 'gawk -W posix -f tregex.awk' ]
#
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# tregex.awk
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
BEGIN{
print _=sprintf("%-20s%-10s%-10s%-10s%-10s\n","regex","input","expected","actual","bug?")
OFS="-"
$(length(_)+1)=""
print $0
while(getline <ARGV[1]) # ADR: was testre.dat
{
RE=$1;IN=$2;OUT=$3
$0=IN
gsub(RE,"x")
printf "%-20s%-10s%-10s%-10s%-10s\n",RE,IN,OUT,$0,$0==OUT?"":"**BUG**"
}
}
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#
# This is the test data file used;
#
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# testre.dat
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# (^) 12345 x12345
# ($) 12345 12345x
# (^)|($) 12345 x12345x
# ($)|(^) 12345 x12345x
# 2 12345 1x345
# (^)|2 12345 x1x345
# 2|(^) 12345 x1x345
# ($)|2 12345 1x345x
# 2|($) 12345 1x345x
# (2)|(^) 12345 x1x345
# (^)|(2) 12345 x1x345
# (2)|($) 12345 1x345x
# ($)|(2) 12345 1x345x
# ((2)|(^)). 12345 xx45
# ((^)|(2)). 12345 xx45
# .((2)|($)) 12345 x34x
# .(($)|(2)) 12345 x34x
# (^)|6 12345 x12345
# 6|(^) 12345 x12345
# ($)|6 12345 12345x
# 6|($) 12345 12345x
# 2|6|(^) 12345 x1x345
# 2|(^)|6 12345 x1x345
# 6|2|(^) 12345 x1x345
# 6|(^)|2 12345 x1x345
# (^)|6|2 12345 x1x345
# (^)|2|6 12345 x1x345
# 2|6|($) 12345 1x345x
# 2|($)|6 12345 1x345x
# 6|2|($) 12345 1x345x
# 6|($)|2 12345 1x345x
# ($)|6|2 12345 1x345x
# ($)|2|6 12345 1x345x
# 2|4|(^) 12345 x1x3x5
# 2|(^)|4 12345 x1x3x5
# 4|2|(^) 12345 x1x3x5
# 4|(^)|2 12345 x1x3x5
# (^)|4|2 12345 x1x3x5
# (^)|2|4 12345 x1x3x5
# 2|4|($) 12345 1x3x5x
# 2|($)|4 12345 1x3x5x
# 4|2|($) 12345 1x3x5x
# 4|($)|2 12345 1x3x5x
# ($)|4|2 12345 1x3x5x
# ($)|2|4 12345 1x3x5x
# x{0}((2)|(^)) 12345 x1x345
# x{0}((^)|(2)) 12345 x1x345
# x{0}((2)|($)) 12345 1x345x
# x{0}(($)|(2)) 12345 1x345x
# x*((2)|(^)) 12345 x1x345
# x*((^)|(2)) 12345 x1x345
# x*((2)|($)) 12345 1x345x
# x*(($)|(2)) 12345 1x345x
# x{0}^ 12345 x12345
# x{0}$ 12345 12345x
# (x{0}^)|2 12345 x1x345
# (x{0}$)|2 12345 1x345x
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#
# I've attached a full copy of this e-mail in ZIP format
# in case of e-mail transport errors corrupting the data.
#
# I've posted the same bug report to gnu.utils.bug and
# it's being discussed in this thread on comp.lang.awk;
#
# From: laura@madonnaweb.com (laura fairhead)
# Newsgroups: comp.lang.awk
# Subject: bug in gawk3.1.0 regex code
# Date: Wed, 08 May 2002 23:31:40 GMT
# Message-ID: <3cd9b0f7.29675926@NEWS.CIS.DFN.DE>
#
#
# byefrom
#
# Laura Fairhead
#
#
#
#
# --------------------
# talk21 your FREE portable and private address on the net at http://www.talk21.com
# ----GgOuLpDpIyE--1020998322088--
# Content-Type: : application/zip;; Name="COPY.ZIP"
# Content-Transfer-Encoding: base64
# Content-Disposition: attachment; filename="COPY.ZIP"
#
# UEsDBBQAAAAIALoaqiyj8d/bjwMAAKsaAAADAAAARklMrVjfa+JAEH4P5H8ISwrRU9EYfbheKBR6
# xRcLvevbYbFtzsqJlBrpQr3722+zMWZ31pk1MaG0Q/m+nR87O9kvruM6/5p4XOc9WSTc05/l
# +m2bSivhb8lzmrx43vw53c5X2f+etourHOc63XMe1wlmLQ8+g3AYjaTFD2ZplY9g+xRbWly3
# NPastYMrQN9cs4DvHYz+dHbomY8SOTctGDlcQfXND1Uz6cK3EXcVdpY37ltSuB55u339cNtu
# F76NPTudHYR0zS2RZ/sd1maHVLdYI/cp31b2PvFW72jkvIi2tLTI94nXY/eCfeZK8Ap7GO1b
# u7QAO8+8FjsLfFx7OowtfW6dLYRv22wZ031uYYc7M/aK5xvEfjp7vDPnQxW2OZuqndDxWeyw
# dt6y5rXPt5xrqG8bW9a8tm8ZN1q1UyYTXvNT2HjN7VWLLL3GR7pl9nlUkx1Z+5xm2/qcYsu4
# z2KHtfOWNad6jR92jGN9jvm2sSNbn1vYlj4n2TLus9h4zW1s/tn/e3iHV55MOXumvUarsvVX
# +OknNGfrr/AK7DbMulLkbZh1VTa8uFSLHF5cqlVt5tW9eWRsH2VbVY10rp+TCu9Q6Rxj2/Ju
# SJE2KG5TqW57848/jS15fXM7mX66ztv7cp16j/FGGr8DdtEN+5uL7sD49WvNOkwGIv5KaS3+
# FsJamLmyFkYmrFnLde6+/4hZl7mOH6yS9SJ9DR5bXwatmLHCrd/PivTxulwlwSJJV8t14n1j
# abIRCfde5mm2iojx/ib2B5eTaeyHl3cPP2N/KNbsx5Op6yw226fg/qbDeIbNc/DoHAR6Mu2I
# dTp+X/zEsTCvGPvK9j0govsrfxqqdJN9cKhMY0vilwdPOebmRwqIy4+x+Tni+Hrc/PKAAnGZ
# 7pXH2fyaYK6X4+B9CcPBt/RRt9z8FoDhoOpH/QJ9j+KAkkf9As2O4oA6N/xy6RWo8OMoqLYN
# 1DDipqo+joIqEGtQqDWJRibXK9oO6igMB1Uu2XeKZwwHlSuO0zue6idVGVE4VQPheeiVIc8F
# sV6Bg6oRx+knkup3Kl8VR+Vb5qGru2N14SNTx2E4qNhwnH1/+chUYRROvfvjeejK6khdeLm/
# +HoFDqolHGfdX17sG5WviqPyLXBQ1WB9D/ULjSvHH9ZXUJOgOKA+UL9AZ1A4dThTftXxTOWh
# qgRs7kI9gF4gwM0fnVfgjo/F19A96T9QSwECFAAUAAAACAC6Gqoso/Hf248DAACrGgAAAwAA
# AAAAAAABACAAAAAAAAAARklMUEsFBgAAAAABAAEAMQAAALADAAAAAA==
# ----GgOuLpDpIyE--1020998322088----
#
#
#