Donna Garverick writes:

> we've got a job that uses grep that *occasionally* gives us
> strange results.
...
> every so often, the file produced by
> grep will have a 'split' record.  that is, some portion of
> an input record (which looks perfectly normal) will end up
> divided onto two records on the output file.

Yep; that's familiar. :(

> while i'm not
> ruling out an escape character of some sort on the input
> record, i think it's quite unlikely.

Given my prior experience, I think we can rule this out.

> so, i'm wondering if
> anyone has run into grep doing 'strange' things like
> this?            - d

Not grep -- my problem was with cousin sed. ;)

I remember only that the problem occurred, not the details
thereof.  Fortunately, hp3000-L has archives. :)

http://raven.utc.edu/cgi-bin/WA.EXE?A2=ind9812B&L=hp3000-l&P=R2158

I'm afraid I cannot offer more than what's there, but hopefully
that will set you on the path of a workaround.

--Glenn