From:	SMTP%"lionel@quark.enet.dec.com"  8-SEP-1994 09:08:46.90
To:	EVERHART
CC:	
Subj:	Re: XAB and FAB help required

From: lionel@quark.enet.dec.com (Steve Lionel)
X-Newsgroups: comp.os.vms
Subject: Re: XAB and FAB help required
Date: 7 Sep 1994 23:08:48 GMT
Organization: Digital Equipment Corporation - Nashua, NH
Lines: 61
Distribution: world
Message-Id: <34lh60$325@jac.zko.dec.com>
Reply-To: lionel@quark.enet.dec.com
Nntp-Posting-Host: quark.zko.dec.com
To: Info-VAX@Mvb.Saic.Com
X-Gateway-Source-Info: USENET

In article <34k32o$sg0@unicorn.ccc.nottingham.ac.uk>, 
cczjtw@unicorn.nott.ac.uk (John Whitehorn) writes:
>Today I needed to modify a program that had been working, untouched, since 1991,
>and guess what, when I complied it, it failed on some of the original coding. Its
>VAX fortran. I don't remember how I wrote the code in the first place, I think
>the brain cells involved have long since been overwritten. The code is a function
>called from LIB$DELETE_FILE as the confirm_routine to set the protection on a
>file (s:werd,o:werd,g:werd,w:werd), its used to remove a users filestore when the
>username is removed.
>
>
>      include '($fabdef)'
>      include '($xabdef)'
>      include '($xabprodef1)'
>      record / fabdef / fab
>      record / xabprodef1 / xabpro
>      .
>      .
>      fab.fab$b_fac=fab$m_upd
>      fab.fab$l_xab=%loc(xabpro)
>      xabpro.xabprodef$$_fill_1=xab$c_pro
>      xabpro.xabprodef$$_fill_2=xab$c_prolen
>
>well guess what DEC no longer have fields called xabprodef$$_fill_1 and 2. does
>anyone have any idea what I need to change them to?

Here's what I recommend.

	include '($xabdef)'
	include '($xabprodef)'
	structure /my_xabprodef/
	  union
	    map
	    record /xabdef/ header
	    end map
	    map
	    record /xabprodef/ body
	    end map
	  end union
	end structure
	record /my_xabprodef/ xabpro
	xabpro.header.xab$b_cod = xab$c_pro
	xabpro.body.[whatever] = something


The $$ fields were never documented and were subject to change at the
whim of the SDL back-end.  Unforunately, RMS defines the XABs in two
pieces - this is fine for "flat structure" languages such as MACRO
and BLISS but causes havoc for everyone else.  Pascal and Ada have
hand-edited definitions (not sure about C) - I prefer to avoid these
so as to not get caught when VMS changes something.  The technique
above may look a bit clunky but it will work and won't break in the
future.

Steve Lionel                      Mail: lionel@quark.enet.dec.com
DEC Fortran Development           WWW:  http://www.digital.com/info/slionel.html
Digital Equipment Corporation     
110 Spit Brook Road, ZKO2-3/N30
Nashua, NH 03062-2698             "Free advice is worth every cent"

For a summary of Digital's Internet services, send mail to info@digital.com