> Not sure why, but we've seen this many times...Usually
> doing the first
A reply:
> :LISTF is bogus, but following ones are accurate; Seems like
> something it's flushed until (After) you look at it.
Original author:
> I have a message file defined as follows
...
> LISTF,2 reports a non zero EOF when I know it is empty.
...
> Then (bizarrely) if exit SH and redo LISTF,2 the EOF is 0 (correct !)

Pre- LISTF,access days (5.5 PP 5 and earlier) the LISTF code got the EOF
from the file label, which is not updated until the file is closed.
Back then the LISTF code actually pseudo opened each file, got all the
needed info and pseudo closed the file. So the first LISTF got (stale)
info prior to the file being closed, and then closed the file.  The next
LISTF got current info, since the file had just been closed by the first
LISTF.  LISTF no longer works this way after 5.5 PP5.  It goes to in
memory data structures and computes an accurate EOF.

regards,
Jeff