It's deja vu all over again! I was trying to generate, and time, some examples of the problem with
"do while lines(ifn ift ifm ) /= 0" when the results began to look surprisingly familiar. I eventually remembered that I had posted something to VMSHARE about REXX level 2 I/O, so I went looking for it. I found I had posted a performance problem with lines() that became APAR VM60551 POOR PERFORMANCE FOR REXX STREAM I/O WITH SFS FILES. Now i'm wondering if I've hit the same bug again, or a related bug, or if this is just "working as designed". See PROB REXX beginning with "Append on 04/17/96 at 03:13 by Alan Ackerman" at <http://www2.marist.edu/~vmshare/browse?fn=REXX&ft=PROB#1923>. PROB REXX reminded me of Kent Fiala's TRACEXEC. (Thank, Kent!) I'm trying to narrow down my performance questions using TRACEXEC. (TRACEXEC is available on the VM Downloads page.)
