[Info-vax] RMS internals?

[P. Sture]

In article <howard-7A9971.14031208082009 at news.newsguy.com>,
 Howard S Shubs <howard at shubs.net> wrote:

> In article <4a7dbb58$0$295$14726298 at news.sunsite.dk>,
>  Arne Vajhøj <arne at vajhoej.dk> wrote:
> 
> > In that case it is not a really a valid file.
> 
> I'm not sure how you arrive at this result.  A file can have any record.  
> RMS asserts limits which other operating systems do not have to adhere 
> to.  So I've written code which "deals" with RMS' limits.
> 
> 
> > See:
> >    http://h71000.www7.hp.com/doc/731final/4523/4523pro_006.html#fab_mrs
> 
> Yes, BTDT.  Those are RMS-specific limits.  If the file had been 
> generated on VMS, maybe they'd be relevant.  Since the file was 
> generated elsewhere, they're not.  This is a case of VMS being "too 
> nice" and getting in my way.  It happens on occasion.

Before VMS COBOL had the ability to produce non-VFC REPORT files, it 
used the fixed part of the record to insert line and form feeds.  This 
was fine if you were outputting to a VMS printer, but not if you wanted 
to transfer the file to a PC and print from there.

I did a similar exercise with those files in the early 1990s, reading by 
block and unpacking the records myself.  IIRC the only problem I ran 
into was that I hadn't realised that records with an odd number of bytes 
were padded with a null filler byte, presumably so that each record 
started on a word boundary.

-- 
Paul Sture