1# From beebe@math.utah.edu Thu Aug 2 15:35:07 2001 2# Received: from mail.actcom.co.il [192.114.47.13] 3# by localhost with POP3 (fetchmail-5.7.4) 4# for arnold@localhost (single-drop); Thu, 02 Aug 2001 15:35:07 +0300 (IDT) 5# Received: by actcom.co.il (mbox arobbins) 6# (with Cubic Circle's cucipop (v1.31 1998/05/13) Thu Aug 2 16:02:36 2001) 7# X-From_: beebe@sunshine.math.utah.edu Thu Aug 2 15:41:13 2001 8# Received: from lmail.actcom.co.il by actcom.co.il with ESMTP 9# (8.9.1a/actcom-0.2) id PAA01349 for <arobbins@actcom.co.il>; 10# Thu, 2 Aug 2001 15:41:06 +0300 (EET DST) 11# (rfc931-sender: mail.actcom.co.il [192.114.47.13]) 12# Received: from billohost.com (www.billohost.com [209.196.35.10]) 13# by lmail.actcom.co.il (8.11.2/8.11.2) with ESMTP id f72Cf3I21032 14# for <arobbins@actcom.co.il>; Thu, 2 Aug 2001 15:41:05 +0300 15# Received: from fencepost.gnu.org (we-refuse-to-spy-on-our-users@fencepost.gnu.org [199.232.76.164]) 16# by billohost.com (8.9.3/8.9.3) with ESMTP id IAA28585 17# for <arnold@skeeve.com>; Thu, 2 Aug 2001 08:34:38 -0400 18# Received: from sunshine.math.utah.edu ([128.110.198.2]) 19# by fencepost.gnu.org with esmtp (Exim 3.22 #1 (Debian)) 20# id 15SHjG-00036x-00 21# for <arnold@gnu.org>; Thu, 02 Aug 2001 08:37:30 -0400 22# Received: from suncore.math.utah.edu (IDENT:GsUbUdUYCtFLRE4HvnnvhN4JsjooYcfR@suncore0.math.utah.edu [128.110.198.5]) 23# by sunshine.math.utah.edu (8.9.3/8.9.3) with ESMTP id GAA00190; 24# Thu, 2 Aug 2001 06:37:04 -0600 (MDT) 25# Received: (from beebe@localhost) 26# by suncore.math.utah.edu (8.9.3/8.9.3) id GAA20469; 27# Thu, 2 Aug 2001 06:37:03 -0600 (MDT) 28# Date: Thu, 2 Aug 2001 06:37:03 -0600 (MDT) 29# From: "Nelson H. F. Beebe" <beebe@math.utah.edu> 30# To: arnold@gnu.org 31# Cc: beebe@math.utah.edu 32# X-US-Mail: "Center for Scientific Computing, Department of Mathematics, 322 33# INSCC, University of Utah, 155 S 1400 E RM 233, Salt Lake City, UT 34# 84112-0090, USA" 35# X-Telephone: +1 801 581 5254 36# X-FAX: +1 801 585 1640, +1 801 581 4148 37# X-URL: http://www.math.utah.edu/~beebe 38# Subject: awk implementations: a bug, or new dark corner? 39# Message-ID: <CMM.0.92.0.996755823.beebe@suncore.math.utah.edu> 40# Status: RO 41# 42# Consider the following program: 43# 44# % cat bug.awk 45BEGIN { 46 split("00/00/00",mdy,"/") 47 if ((mdy[1] == 0) && (mdy[2] == 0) && (mdy[3] == 0)) 48 { 49 print "OK: zero strings compare equal to number zero" 50 exit(0) 51 } 52 else 53 { 54 print "ERROR: zero strings compare unequal to number zero" 55 exit(1) 56 } 57} 58# 59# Here are the awk implementation versions (on Sun Solaris 2.7): 60# 61# % awk -V 62# awk version 19990416 63# 64# % mawk -W version 65# mawk 1.3.3 Nov 1996, Copyright (C) Michael D. Brennan 66# 67# % nawk -V 68# awk version 20001115 69# 70# % gawk --version 71# GNU Awk 3.1.10 72# ... 73# 74# Here's what they say about the test program: 75# 76# foreach f (awk mawk nawk gawk gawk-*) 77# echo ======== $f 78# $f -f ~/bug.awk 79# end 80# 81# ======== awk 82# OK: zero strings compare equal to number zero 83# ======== mawk 84# OK: zero strings compare equal to number zero 85# ======== nawk 86# OK: zero strings compare equal to number zero 87# ======== gawk 88# ERROR: zero strings compare unequal to number zero 89# ======== gawk-3.0.0 90# OK: zero strings compare equal to number zero 91# ======== gawk-3.0.1 92# OK: zero strings compare equal to number zero 93# ======== gawk-3.0.3 94# OK: zero strings compare equal to number zero 95# ======== gawk-3.0.4 96# OK: zero strings compare equal to number zero 97# ======== gawk-3.0.5 98# OK: zero strings compare equal to number zero 99# ======== gawk-3.0.6 100# OK: zero strings compare equal to number zero 101# ======== gawk-3.0.60 102# OK: zero strings compare equal to number zero 103# ======== gawk-3.0.90 104# ERROR: zero strings compare unequal to number zero 105# ======== gawk-3.0.91 106# ERROR: zero strings compare unequal to number zero 107# ======== gawk-3.0.92 108# ERROR: zero strings compare unequal to number zero 109# ======== gawk-3.0.93 110# ERROR: zero strings compare unequal to number zero 111# ======== gawk-3.0.94 112# ERROR: zero strings compare unequal to number zero 113# ======== gawk-3.0.95 114# ERROR: zero strings compare unequal to number zero 115# ======== gawk-3.0.96 116# ERROR: zero strings compare unequal to number zero 117# ======== gawk-3.0.97 118# ERROR: zero strings compare unequal to number zero 119# ======== gawk-3.1.0 120# ERROR: zero strings compare unequal to number zero 121# ======== gawk-3.1.10 122# ERROR: zero strings compare unequal to number zero 123# 124# Identical results were obtained on Apple Rhapsody, Apple Darwin, 125# Compaq/DEC Alpha OSF/1, Intel x86 GNU/Linux, SGI IRIX 6.5, DEC Alpha 126# GNU/Linux, and Sun SPARC GNU/Linux, so it definitely is not a C 127# compiler problem. 128# 129# However, the gray awk book, p. 44, says: 130# 131# In a comparison expression like: 132# x == y 133# if both operands have a numeric type, the comparison is numeric; 134# otherwise, any numeric operand is converted to a string and the 135# comparison is made on the string values. 136# 137# and the new green gawk book, p. 95, says: 138# 139# When comparing operands of mixed types, numeric operands are 140# converted to strings using the value of `CONVFMT' 141# 142# This suggests that the OK response in bug.awk is wrong, and the ERROR 143# response is correct. Only recent gawk releases do the right thing, 144# and it is awk, mawk, and nawk that have a bug. 145# 146# If I change the test program from "00/00/00" to "0/0/0", all versions 147# tested produce the OK response. 148# 149# Comments? 150# 151# After reading the two book excerpts, I changed my code to read 152# 153# if (((0 + mdy[1]) == 0) && ((0 + mdy[2]) == 0) && ((0 + mdy[3]) == 0)) 154# 155# and output from all implementations now agrees. 156# 157# ------------------------------------------------------------------------------- 158# - Nelson H. F. Beebe Tel: +1 801 581 5254 - 159# - Center for Scientific Computing FAX: +1 801 585 1640, +1 801 581 4148 - 160# - University of Utah Internet e-mail: beebe@math.utah.edu - 161# - Department of Mathematics, 322 INSCC beebe@acm.org beebe@computer.org - 162# - 155 S 1400 E RM 233 beebe@ieee.org - 163# - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe - 164# ------------------------------------------------------------------------------- 165# 166