Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!think!bloom-beacon!mit-eddie!ll-xn!cit-vax!ucla-cs!zen!ucbvax!LBL.ARPA!nagy%warner.hepnet
From: nagy%warner.hepnet@LBL.ARPA
Newsgroups: comp.os.vms
Subject: RE: VAX C Linker (really Librarian) problems
Message-ID: <870910061656.024@Lbl.Arpa>
Date: Thu, 10-Sep-87 09:16:55 EDT
Article-I.D.: Lbl.870910061656.024
Posted: Thu Sep 10 09:16:55 1987
Date-Received: Sat, 12-Sep-87 10:20:04 EDT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 21

>Jerry's reply re the Vax C Linker Bug was very welcome, as we
>have had it occur in linking some large programs. The programs
>always run fine.. is it likely that the error in PSECT sizes
>is likely to cause crashes, or obscure data-dependent errors ?
     
Yes, remember that PSECTs are just like FORTRAN COMMON's (in fact
that's how FORTRAN COMMON's are handled under VMS, each is a
separate PSECT).  The biggest problem manifests itself in linking
with VAXCRTL.EXE when VAXCRTL is placed in an image library - the
Librarian doesn't keep the necessary information around and you
apparently end up with your program thinking some extern variable
(PSECT) is one size and VAXCRTL routines thinking its another...


     

= Frank J. Nagy   "VAX Guru"
= Fermilab Research Division EED/Controls
= HEPNET: WARNER::NAGY (43198::NAGY) or FNAL::NAGY (43009::NAGY)
= BitNet: NAGY@FNAL
= USnail: Fermilab POB 500 MS/220 Batavia, IL 60510