bin/59903: rs dumps core on valid input

["mac%culver.net@localhost via gnats" <gnats-admin%NetBSD.org@localhost>]

>Number:         59903
>Category:       bin
>Synopsis:       rs dumps core on valid input
>Confidential:   no
>Severity:       serious
>Priority:       low
>Responsible:    bin-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jan 10 01:00:00 +0000 2026
>Originator:     mac%culver.net@localhost
>Release:        NetBSD 10.1
>Organization:
>Environment:
System: NetBSD SS.Culver.Net 10.1 NetBSD 10.1 (MIKE-$Revision: 2025-Dec-27 $) #1: Sun Dec 28 19:39:48 UTC 2025 mac%SS.Culver.Net@localhost:/usr/src/sys/arch/amd64/compile/MIKE amd64   My kernel is just a rename of GENERIC
Architecture: x86_64
Machine: amd64
>Description:
rs appears to work properly most of the time, but by accident I found the following error
>How-To-Repeat:
file '2x2' contains:
43   76
27   29

file '2x10' contains"
43   27   8    22   12   78   18   43   41   13
76   29   85   100  34   78   45   80   24   50

$ cat 2x2
43   76
27   29

$ rs -tz < 2x2
43  76  27  29

$ cat 2x10
43   27   8    22   12   78   18   43   41   13
76   29   85   100  34   78   45   80   24   50

$ rs -tz < 2x10
zsh: segmentation fault  rs -tz < 2x10

I did not try reducing the seg fault input more than this 2x10 example


test files generated by playing around with this:

$ cd /tmp
$ jot -r 100 | rs 10 10 | tee 10x10 | rs -T > T10x10
$ head -2 T10x10 > 2x10
$ rs < 2x10 | rs | head -4 | rs 2 2  > 2x2
$ rs -tz < 2x10
zsh: segmentation fault  rs -tz < 2x10


I believe the problem is specific input data, to wit:

$ cat 2x10
43   27   8    22   12   78   18   43   41   13
76   29   85   100  34   78   45   80   24   50

$ rs -z -t < 2x10
zsh: segmentation fault (core dumped)  rs -z -t < 2x10     ;; switch order is irrelevant

$ cat /tmp/2x10
97  2   2   29  40  6   67  36  54  78
88  14  26  77  10  85  9   65  96  97

$ rs -t -z < /tmp/2x10
97   2   2  29  40   6  67  36  54  78  88  14  26  77  10  85   9  65  96  97


FINALLY, if I change the "100" in the 'bad' file to, say, 99 --- no problem.  to 101?  same problem.

>Fix: