agrep gives incorrect output in some approximate searches

[Eric Price]

agrep sometimes finds a line under exact matching, but not under inexact matching. I couldn't come up with a small test case, but
http://web.mit.edu/ecprice/Public/broken-agrep-example.gz has a 1.3 million line test case:

$ wget http://web.mit.edu/ecprice/Public/broken-agrep-example.gz -O - | zcat > case
[...]
$ agrep ^waterfal$ case #correct
waterfal
$ agrep -S1 ^waterfal$ case #should be identical to previous, but isn't
$ agrep -1 ^waterfal$ case #incorrect in last line
waterfall
waterval
aterfal
$ agrep ^aterfal$ case #as you can see here.
$

I'm also a bit confused by the relation between the version in Ubuntu and the version (3.41) at http://www.tgries.de/agrep/ . The ubuntu version number is 4.17, but agrep -V claims to be version 3.0. It seems like version 3.41 had a substantial amount of development in 96-97. It also doesn't have this bug.

ProblemType: Bug
DistroRelease: Ubuntu 11.10
Package: agrep 4.17-7
ProcVersionSignature: Ubuntu 3.0.0-14.23-generic 3.0.9
Uname: Linux 3.0.0-14-generic x86_64
NonfreeKernelModules: openafs
ApportVersion: 1.23-0ubuntu4
Architecture: amd64
Date: Mon Jan 9 18:47:11 2012
InstallationMedia: Ubuntu 11.04 "Natty Narwhal" - Beta amd64 (20110413)
ProcEnviron:
 LC_COLLATE=C
 PATH=(custom, user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: agrep
UpgradeStatus: Upgraded to oneiric on 2011-10-14 (87 days ago)

[Dave Gilbert (ubuntu-treblig)]

Hi Eric,
  Hmm, yes this is a bit confusing,

I agree your 'agrep -S1' case is producing the wrong result, however in 4.17-9 the manpage says:
' -Sk Set the cost of a substitution to k (k is a positive integer). This option does not currently work with regular
              expressions.'

so, I think that's just a limitation on agrep 4.17

I agree the -1 case is wrong to emit 'aterfal' since I agree it's nowhere in the input (and that's why I've confirmed it).

As for versions, yes, it's a mess, the 'homepage' listed in debian/control points to http://freshmeat.net/projects/agrep whcih points to the 2.04 version ftp://ftp.cs.arizona.edu/agrep/

The package has a 4.17 source tar and from what I can see it comes from http://wftp.tu-chemnitz.de/pub/Local/urz/ding/contrib/SRPMS/agrep-4.17-1.src.rpm that was suggested in the following debian bug

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=307442

but as you say the printed version number is completely inconsistent in that source.