[net.bugs.4bsd] A loop in 4.2 f77 do_fio

kel@ea.UUCP (05/29/84)

#N:ea:1700003:000:831
ea!kel    May 29 14:45:00 1984

As I have gotten some inquiries after my admittedly inadequate description
of a forever looping condition in the 4.2bsd do_fio, here's more:

The bug I was bit by is evoked by formats of the form 'a0', 'i0', etc.
for example:

      character * 80 thing
      data thing/'This is a test thing.  This has been a test thing.'/
      write(6, 1001) thing
1001  format(a0)
      stop
      end

	will write output to stdout forever.

As will be apparent by now, this problem is an artifact of a very
obscure dialect (bastardization?) of fortran.  By the way, when I
cooked up this example, I discovered that the formatter detects
an invalid format if you put the 'a0' directly into the write
statement.

If you get a fix for this, it might be nice for me to have, though
I've been warned about it, since I don't have source.

				Ken