#From hankedr@manatee.dms.auburn.edu  Tue Oct 13 22:15:59 1998

#Return-Path: <hankedr@manatee.dms.auburn.edu>

#Received: from cssun.mathcs.emory.edu (cssun.mathcs.emory.edu [170.140.150.1])

#	by dmx.netvision.net.il (8.9.0.Beta5/8.8.6) with ESMTP id PAA03924

#	for <arobbins@netvision.net.il>; Tue, 13 Oct 1998 15:32:13 +0200 (IST)

#Received: from mescaline.gnu.org (we-refuse-to-spy-on-our-users@mescaline.gnu.org [158.121.106.21]) by cssun.mathcs.emory.edu (8.7.5/8.6.9-940818.01cssun) with ESMTP id KAA11644 for <arnold@mathcs.emory.edu>; Tue, 13 Oct 1998 10:22:32 -0400 (EDT)

#Received: from manatee.dms.auburn.edu (manatee.dms.auburn.edu [131.204.53.104])

#	by mescaline.gnu.org (8.9.1a/8.9.1) with ESMTP id KAA03250

#	for <arnold@gnu.org>; Tue, 13 Oct 1998 10:25:32 -0400

#Received: (from hankedr@localhost)

#	by manatee.dms.auburn.edu (8.9.1a/8.9.1) id JAA13348;

#	Tue, 13 Oct 1998 09:22:29 -0500 (CDT)

#Date: Tue, 13 Oct 1998 09:22:29 -0500 (CDT)

#Message-Id: <199810131422.JAA13348@manatee.dms.auburn.edu>

#From: Darrel Hankerson <hankedr@dms.auburn.edu>

#To: arnold@gnu.org

#In-reply-to: <199810131313.QAA31784@alpha.netvision.net.il> (message from

#	Aharon Robbins on Tue, 13 Oct 1998 16:10:36 +0200)

#Subject: Re: full text of bug report?

#Mime-Version: 1.0

#Content-Type: text/plain; charset=US-ASCII

#X-UIDL: bf3fce492dad4ab030c561e7b2f27d0a

#Status: RO

#

#   Do you have the full text of the	a = a "\n" f() 	bug report?

#   I can't find it.... I'm not sure there really is a bug.

#

#Yes, see below.  

#

#His example has unnecessary fragments (in particular, the use of

#gensub is irrelevant).  As I wrote to you earlier, the interesting

#question for me is:

#

#   Is the concatenation result undefined?  If the result is defined or

#   implementation-dependent, then gawk has a bug.

#

#

#=== Original report =====================================================

#From: Attila Torcsvari <arcdev@mail.matav.hu>

#To: "'bug-gnu-utils@prep.ai.mit.edu'" <bug-gnu-utils@gnu.org>

#Subject: gawk 3.0.3 bug

#Date: Thu, 17 Sep 1998 18:12:13 +0200

#MIME-Version: 1.0

#Content-Transfer-Encoding: 7bit

#Resent-From: bug-gnu-utils@gnu.org

#X-Mailing-List: <bug-gnu-utils@gnu.org> archive/latest/3396

#X-Loop: bug-gnu-utils@gnu.org

#Precedence: list

#Resent-Sender: bug-gnu-utils-request@gnu.org

#Content-Transfer-Encoding: 7bit

#Content-Type: text/plain; charset="us-ascii"

#Content-Length: 618

#

#Bug-gnuers,

#please pass it to the responsible.

#

#The following generates something interesting:

#

BEGIN{

a="aaaaa"

a=a a #10

a=a a #20

a=a a #40

a=a a #80

a=a a #160

a=a a # i.e. a is long enough

a=a"\n"f() # this causes the trouble

print a # guess the result

}

function f()

{

#print "a before: ", a

#a=gensub("a","123,","g",a) # 'a' will be just a bit longer (4 times, but still should fit: 4*160=640)

gsub(/a/, "123", a)

#print "a after: ", a

return "X"

}

#

#Possible reason:

#during f the a is modified,

#it can be even freed, because gensub modifies its size

#the printout contains trash.

#

#Used version: VC compiled WinNT 32 bit Intel.

#

#Regards,

#

#Attila Torcsvari

#Arcanum Development

#