RE: [open-cobol-list] Clarification requested (LINAGE - LS - NL)

[Roger While]

Thanks Bill, That helps a little.
Yes, I know that LS is an MF extension. IMHO, it should
have been adopted into the standard. It makes a lot of sense.
This is what I am going to do :
If the file is (LINE) SEQUENTIAL and has LINAGE,
(it is checked that the LINAGE applies to a SEQ file)
it will internally operate as LINE SEQ.

 Another question which I think is logically clear but not
specifically answered in any doc although I think it may be
implied :
Assume -
LINAGE data-item = <some number>
FOOT data-item = 0
TOP data-item = 0
BOTTOM data-item = 0

We OPEN OUTPUT the file and start WRITE'ing.
At some point we MOVE 1 to TOP and BOTTOM.
We now get an END-OF-PAGE condition on
a WRITE BEFORE/AFTER 3.
At this point we reread the 4 values above.
It's pretty obvious that we should use the previous
value of BOTTOM (plus previous LINAGE minus previous
LINAGE-COUNTER) to finish off the current page.
But I assume that TOP value (for the next page)
to be used should be the new value.
Correct ?

Roger

> As usual, things are a "little" complicated <G>
>
> First and foremost, "line sequential" is NOT ANSI or ISO conforming ('85 
> or
> '02 Standards - or any other Standard).  Therefore, there is NOTHING in 
> any
> Standard telling you what to do with WRITE or anything else of LINE
> sequential files.
>
> In fact, the current N/E V4 Micro Focus LRM gives different rules for 
> "line
> sequential" (marked with an "MF" bubble) and record (ANSI conforming)
> sequential files.  They are:
>
> GR(10) (line sequential) - If the ADVANCING phrase is not used, automatic
> advancing of one line is provided to act in accordance with the convention
> of your operating system text editor (usually as if you had specified 
> BEFORE
> ADVANCING 1 LINE).
>
>    ***
>
> GR(15 (record sequential) - If the ADVANCING phrase is not used, automatic
> advancing is provided when output is directed to a list device (PRINTER or
> PRINTER-1), to act as if you had specified AFTER ADVANCING 1 LINE.
>
>    ***
>
> Now the '02 Standard (tried to) clarify what "print type" files and 
> devices
> are/were - but it is pretty clear
>
>   GR(21) - "If the ADVANCING phrase is not used, automatic advancing 
> shall
> be provided by the implementor to act as if the user has specified AFTER
> ADVANCING 1 LINE."
>
>    ***
>
> Given all of this, I think it is "reasonable" to treat LINE SEQUENTIAL 
> files
> with "BEFORE ADVANCING" and RECORD SEQUENTIAL files with "AFTER 
> ADVANCING".
>
>
> Now what to do with output sent directly to a "printer-type" device is not
> nearly so clear.  On the other hand, I think that using the Select/Assign
> "organization" to determine BEFORE/AFTER advancing is reasonable - even 
> for
> these.
>
> Does this help?
>
> > -----Original Message-----
> > From: address@hidden
> > [mailto:address@hidden On
> > Behalf Of Roger While
> > Sent: Sunday, February 27, 2005 2:33 PM
> > To: address@hidden
> > Subject: [open-cobol-list] Clarification requested (LINAGE - LS - NL)
> >
> > After preparing most of the code for implementation of
> > LINAGE, I noticed the following and would appreciate comments :
> > (Bill ?)
> > 1) A WRITE to a (LINE) SEQUENTIAL file without an ADVANCING is
> >       a) 2002 says it is AFTER ADVANCING 1
> >       b) MF and OpenCobol says it is BEFORE ADVANCING 1
> >
> > Given (1) and (a),  if this is the first line to be printed,
> > should the first line to be produced consist of only a
> > "new line" character ? Or is "top of page" special ?
> > (MF produces a single newline)
> > (And with LINAGE MF still produces (IMHO wrongly) a
> >    newline)
> > What about WRITE AFTER PAGE ?
> > Should the LINAGE clause affect this behaviour ?
> >
> > If we do a "WRITE AFTER"  as the last statement before closing
> > the file, we produce (also MF):
> > "<NL>text"
> > and then close the file, we end up with a non-newline terminated
> > output.
> > Should a CLOSE detect this and write a terminating newline ?
> >
> > Should it be possible to "concatenate" a print line by
> > doing a WRITE AFTER followed by a WRITE BEFORE ?
> >
> > (Before you ask Bill, yes, LINAGE-COUNTER will be implemented per
> > file  :-)  )
> >
> > Roger