LSTOWN-L Archives

LISTSERV List Owners' Forum

LSTOWN-L

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Topic: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Ben Parker <[log in to unmask]>
Fri, 2 Nov 2001 20:44:39 -0700
text/plain (37 lines)
On Fri, 2 Nov 2001 20:33:44 EST, Bill Gruber <[log in to unmask]> wrote:

>Now I'm really confused.  The SizeLim keyword takes lines as the argument.
>So does Digest.  And what about Long-Lines -- that works on a line by line
>basis even.

You take a bytestream (i.e. a 'message' of arbitrary length) and store it in a
file system where every record is always exactly 80 char long.   You fill every
byte of filespace with every byte character of the 'message' (treating space and
CRLF, etc. exactly the same as any other byte).  Depending on how it comes out,
you may need to pad the final record to fill out to exactly 80.  Once stored in
such a file system, one way of looking at the 'size' of that 'message' is as a
certain number 'n' of these 80char records or 'lines'.  This is one definition
of what you might call a 'line'.  Remember, since all the storage records are
guaranteed to be the same fixed length, no marker is needed for the
end-of-record or 'end-of-line'.

The 'lines' in question refer only to the number of storage records consumed and
have nothing whatever to do with the count of some number 'L' of arbitrary
length character strings that just happen to be terminated with the special byte
pair of 'CRLF'.  This is a separate definition of a 'line'.

Remember LISTSERV's VM heritage and the file systems that might have been used
on the old mainframes (before my time).  Given the above it should be obvious
that LISTSERV's use of 'lines' for determining SIZELIM and Digest length, may
have almost no resemblance to the number of strings ending in CRLF.  You can
also see that by handling the 'message' this way there is no possibility for
additional arbitrary new CRLF byte pairs ('line breaks') to be inserted anywhere
in the bytestream.

I guess I was incorrect to say LISTSERV has no concept of 'line'.  A better way
to say it would be that LISTSERV's concept of 'line' bears no resemblance to
what we commonly think of as 'lines' in a piece of text.  So it is impossible
for LISTSERV to make arbitrary changes to those structures we think of as
'lines', while at the same time it is also possible to set 'size limits'
expressed in terms of LISTSERV's concept of 'lines'.

ATOM RSS1 RSS2