[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference turris::pli

Title:PL/I Notes
Moderator:PEROIT::LUCIA
Created:Sat Jan 25 1986
Last Modified:Wed May 14 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:707
Total number of notes:2472

706.0. "Alpha VMS PL/I V4.1 PLI-F-ACCESSIBLE "Handler context ..." by EPS::VANDENHEUVEL (Hein) Mon Mar 31 1997 13:40

Hi,
    The following note appeared in comp.os.vms today.
    Anyone care to answer here and/or there? If you 
    answer here, kindly mail me the response as I do
    not normally follow this conference. I'll stick 
    it in the news group if need be.
    
    Regards,
    		Hein.
    
    
Article: 168810
From: Jason Armistead <otis@magna.com.au.no-spam>
Newsgroups: comp.os.vms
Subject: HELP: DEC (Uniprise) PL/I compiler undocumented message
Date: Mon, 31 Mar 1997 14:33:32 -1000
Organization: Info-Vax<==>Comp.Os.Vms Gateway
 
Our programmers have just moved to Alpha VMS PL/I V4.1 (4.something anyway).
 
The source code of one program causes the compiler to bomb with the
following message
 
PLI-F-ACCESSIBLE (??? I think this is the code) and then the text "Handler
context data is allocated at a negative offset from SP"
 
OK, so I figure this has got to do with the ON ENDFILE and ON UNDEFINEDFILE
statements that are part of the MAIN program block, since the ON construct
constitutes the condition handling part of the code.  Somehow the Stack
Pointer gets mucked up in relation to the condition handler contexts.
 
The ON statements are not very complex, just a BEGIN-END pair with one
assignment statement in it e.g. IOSTAT1 = 2;
 
There are, however, about 14 FILE variables for which ON ENDFILE and ON
UNDEFINEDFILE condition handlers are established, a total of 28 ON handlers.
 
Can anyone shed any light on this problem ?
 
The compiler message IS NOT documented in the manuals (tut, tut, tut
Uniprise), nor in HELP /MESSAGE, which is really disappointing.  It emanates
from the PL/I compiler EXE itself.  And since it's FATAL, I don't get usable
object code until I fix it or work around it.
 
As we MUST get this code working on Alpha, since our VAX is disappearing
soon, help is urgently needed.
 
PS: The VAX is running a version of PL/I about 4 years old (before LMF
existed), I think it's version 2.x from memory.  The source code works great
there.  And no, VESTing the image is not an option, since the code is still
"live"
 
Regards
Jason Armistead
DECUS Australia Pathworks/Internetworking SIG chairman
Check out the DECUS home page http://www.decus.org.au
 
OTIS Engineering Centre
Minto, Australia
Really: otis@magna.com.au - headers munged to avoid spam e-mails
Check out the OTIS home page: http://www.otis.com and discover Odyssey
 
Disclaimer: The views expressed above represent those of the author, and not
of OTIS Elevator Company, and are not meant to endorse or otherwise
represent any goods or services mentioned.
 
T.RTitleUserPersonal
Name
DateLines
706.1SPECXN::DERAMODan D'EramoMon Mar 31 1997 15:0517
	Hein,
        
>PLI-F-ACCESSIBLE (??? I think this is the code) and then the text "Handler
>context data is allocated at a negative offset from SP"
        
        I believe the error message was 
        
%PLIG-F-ASSERTION, handler context data is allocated at a negative
offset from SP.
        
        as described in topics 675 and 698 here.
        
        Unfortunately, none of the reports of this problem that I have
        seen has come with reproducing source code that the CSC can
        forward to Uniprise.
        
        Dan
706.2No solution now, but IPMTedSTP01::STPVNV::stp01::voinovVladimir Voinov, MCS RussiaMon Apr 14 1997 07:387
This case (note 698) had been IPMTed (CFS.47758) and is closed
by the moment, because we were assured by Uniprise, that
the problem will be fixed with the new version of compiler
(hopefully, comimg soon). 

Vladimir