[Info-vax] RMS internals?

[glen herrmannsfeldt]

In comp.os.vms Bob Koehler <koehler at eisner.nospam.encompasserve.org> wrote:
(snip, I wrote)
 
<> TCP is a stream protocol, with no record marks.  The resulting
<> file should be independent of anything in the TCP/IP chain.
 
<   FTP imposes a standard CRLF pair at the end of each line of text,
<   when transfering as a text file.  Of course, TCP doesn't need to
<   know about this.

More specifically an ASCII CRLF pair.  The text will be in ASCII
even if the source and/or destination machine do not use ASCII.

With binary mode, no line terminating characters will be added
or removed.  If you transfer fixed length files from IBM mainframes,
they come through without any record marks, but you know where they
are supposed to be.  (You know the LRECL.)  If you transfer variable
length (VB or VBS) files there are no record marks and you often
don't know where they are supposed to be.  

I believe that there is also an FTP option specifically for 36 bit 
systems, but I don't remember how that works.

For my original comment, I was thinking that the ftp program
was separate from the underlying TCP/IP system, but maybe that
isn't true.  I believe I used multinet about 20 years ago, but
not since then.  

Otherwise, there is always SET FILE/ATTRIBUTES to fix up
some of the problems.

-- glen