Help needed on old VMS Fortran 77 Program 
Author Message
 Help needed on old VMS Fortran 77 Program

Surely, you could acquire CVF or any other commercial compiler of your
choice from one of the excellent vendors who maintain a presence on the
web, in Europe or elsewhere.  But, the VMS extensions are well enough
known that most are either understood or diagnosed as invalid by these
compilers.  I would be interested in seeing a list of those which may
typically escape diagnosis, but I wouldn't have any confidence that CVF
would handle those correctly.  If your code does things like using IF(L)
to test whether L is odd or even you should be changing it to something
less inscrutable.  Send me the code and I will have a look and see if I
can do what you asked.  .zip or .gz or .tgz is fine.

--
Tim Prince

Quote:
> Hi you,
> I need some help on an old VMS fortran77 program.

> The problem is, that it compiles without any errors with Watcom,
> Salford and f77 compilers, but does not run correctly. This might be,
> because there are some extensions the compiler dont know.

> I need someone with DR or CVF compilers, who will compile the source
> and pass the error log or the .exe (if there is any) file back to me.

> Or much better: Let me have a DR compiler for testing.
> It should be someone in Germany or in a country in nothern Europe.
> If there is anyone who is willing to pass me the compiler via email or
> ftp I can promise to pass back a protokol of what has happened to the
> program and how it was used.

> I do not want to violate any licence of any company, but I do not see
> any chance to come to a solution in that matter. Compaq told me, they
> do not have any Fortran compilers, even if I told them the order
> number of their CVF.

> If there is anybody out who can offer me some help pls. mail.

> thanks
> Jens



Wed, 18 Jun 1902 08:00:00 GMT  
 Help needed on old VMS Fortran 77 Program
Hi you,
I need some help on an old VMS Fortran77 program.

The problem is, that it compiles without any errors with Watcom,
Salford and f77 compilers, but does not run correctly. This might be,
because there are some extensions the compiler dont know.

I need someone with DR or CVF compilers, who will compile the source
and pass the error log or the .exe (if there is any) file back to me.

Or much better: Let me have a DR compiler for testing.
It should be someone in Germany or in a country in nothern Europe.
If there is anyone who is willing to pass me the compiler via email or
ftp I can promise to pass back a protokol of what has happened to the
program and how it was used.

I do not want to violate any licence of any company, but I do not see
any chance to come to a solution in that matter. Compaq told me, they
do not have any Fortran compilers, even if I told them the order
number of their CVF.

If there is anybody out who can offer me some help pls. mail.

thanks
Jens



Wed, 18 Jun 1902 08:00:00 GMT  
 Help needed on old VMS Fortran 77 Program


Quote:
>Hi you,
>I need some help on an old VMS Fortran77 program.

>The problem is, that it compiles without any errors with Watcom,
>Salford and f77 compilers, but does not run correctly. This might be,
>because there are some extensions the compiler dont know.

Well, it might be. But generally a compiler will be unable to compile an
extension it doesn't understand. (Not always; I once had a problem with
some F77 code which did string(1:n) = string(2:n+1) via a function call.
The compiler didn't flag it at compile or run time, but it was illegal
F77 code and didn't work as the author had intended.) Does it also link
without errors? If it's failing to find VMS library functions it should
tell you at link time.

You have Salford; have you tried running with /undef and /debug? This
will tell you if you try to access an uninitialized variable (top of my
list of possibilities for your problems would be that the VMS version
automatically initialises everything to zero, so the code assumes it) or
run off the end of an array.

If you want to buy CVF and can't find a German dealer you could try

work for them.

Hope this helps,

Catherine.
--
Catherine Rees Lay



Wed, 18 Jun 1902 08:00:00 GMT  
 Help needed on old VMS Fortran 77 Program

Quote:

>I do not want to violate any licence of any company, but I do not see
>any chance to come to a solution in that matter. Compaq told me, they
>do not have any Fortran compilers, even if I told them the order
>number of their CVF.

Please see the "Where to buy Compaq Visual Fortran" list at our web
site at http://www.compaq.com/fortran - it lists several resellers in
Germany.  CVF is sold through independent resellers worldwide, which
is why some local Compaq offices may not be able to tell you much
about it.



Fortran Engineering
Compaq Computer Corporation, Nashua NH

Compaq Fortran web site: http://www.compaq.com/fortran
Compaq Fortran Message Board:
   http://forum.compaq.com:80/mb8/system/fortran_login.html



Wed, 18 Jun 1902 08:00:00 GMT  
 Help needed on old VMS Fortran 77 Program
Having seen the code, I would be extremely disappointed in a compiler
which did not produce fatal error messages, even without specification
of additional diagnostics.  There are a few extensions which would have
been supported on VAX/VMS but not by CVF6.1A, and a few pointless
illegal constructs which may have been accepted on VAX/VMS but not by
CVF6.1A, such as use of P formats without a numerical prefix, and a GOTO
which jumps to a CONTINUE which is shared by inner and outer loops.
Yes, it is full of undefined variables.  If you like, I will check the
Salford f77 some time to see what it says.

--
Tim Prince


Quote:


> >Hi you,
> >I need some help on an old VMS Fortran77 program.

> >The problem is, that it compiles without any errors with Watcom,
> >Salford and f77 compilers, but does not run correctly. This might be,
> >because there are some extensions the compiler dont know.

> Well, it might be. But generally a compiler will be unable to compile
an
> extension it doesn't understand. (Not always; I once had a problem
with
> some F77 code which did string(1:n) = string(2:n+1) via a function
call.
> The compiler didn't flag it at compile or run time, but it was illegal
> F77 code and didn't work as the author had intended.) Does it also
link
> without errors? If it's failing to find VMS library functions it
should
> tell you at link time.

> You have Salford; have you tried running with /undef and /debug? This
> will tell you if you try to access an uninitialized variable (top of
my
> list of possibilities for your problems would be that the VMS version
> automatically initialises everything to zero, so the code assumes it)
or
> run off the end of an array.

> If you want to buy CVF and can't find a German dealer you could try

> work for them.

> Hope this helps,

> Catherine.
> --
> Catherine Rees Lay



Wed, 18 Jun 1902 08:00:00 GMT  
 Help needed on old VMS Fortran 77 Program
All the compilers, including Salford f77, CVF6.1A, and others, are
complaining about different numbers of arguments to a subroutine.  From
bitter experience, I know that CVF6.1A would do better to treat this as
a fatal, as it appears to put the stack out of synchronization.

Among the VAX extensions present, which CVF6.1A appears to accept, but
not all others do, are:

Q format specifiers
BYTE declarations
in line compiler option specifications

Of those, the only one which the Lahey compiler rejects is the in line
option.  I don't recall whether format specifiers are checked at compile
time.

--
Tim Prince

Quote:
> Having seen the code, I would be extremely disappointed in a compiler
> which did not produce fatal error messages, even without specification
> of additional diagnostics.  There are a few extensions which would
have
> been supported on VAX/VMS but not by CVF6.1A, and a few pointless
> illegal constructs which may have been accepted on VAX/VMS but not by
> CVF6.1A, such as use of P formats without a numerical prefix, and a
GOTO
> which jumps to a CONTINUE which is shared by inner and outer loops.
> Yes, it is full of undefined variables.  If you like, I will check the
> Salford f77 some time to see what it says.

> --
> Tim Prince




> > >Hi you,
> > >I need some help on an old VMS Fortran77 program.

> > >The problem is, that it compiles without any errors with Watcom,
> > >Salford and f77 compilers, but does not run correctly. This might
be,
> > >because there are some extensions the compiler dont know.

> > Well, it might be. But generally a compiler will be unable to
compile
> an
> > extension it doesn't understand. (Not always; I once had a problem
> with
> > some F77 code which did string(1:n) = string(2:n+1) via a function
> call.
> > The compiler didn't flag it at compile or run time, but it was
illegal
> > F77 code and didn't work as the author had intended.) Does it also
> link
> > without errors? If it's failing to find VMS library functions it
> should
> > tell you at link time.

> > You have Salford; have you tried running with /undef and /debug?
This
> > will tell you if you try to access an uninitialized variable (top of
> my
> > list of possibilities for your problems would be that the VMS
version
> > automatically initialises everything to zero, so the code assumes
it)
> or
> > run off the end of an array.

> > If you want to buy CVF and can't find a German dealer you could try

I
> > work for them.

> > Hope this helps,

> > Catherine.
> > --
> > Catherine Rees Lay



Wed, 18 Jun 1902 08:00:00 GMT  
 Help needed on old VMS Fortran 77 Program
If you have missing or mis-matched arguments, you could use
FPT (http://simcon.uk.com) which flags them, and which will
fix most of them automatically.

The missing arguments cause a stack violation when the
arguments are passed on the stack.  The routine called
unstacks as many arguments as it needs.  If the call site
has too few, too much data is removed from the stack and a
violation occurs, often at a point remote from the call
which caused the problem.  Great fun to debug.  Numeric
arguments used instead of strings (a favourite old PDP/VAX
trick) cause a similar problem because the string arguments
have a length which is put on the stack but the numerics do
not.

Best wishes,

John Collins.

* Sent from AltaVista http://www.altavista.com Where you can also find related Web Pages, Images, Audios, Videos, News, and Shopping.  Smart is Beautiful



Wed, 18 Jun 1902 08:00:00 GMT  
 
 [ 7 post ] 

 Relevant Pages 

1. Help needed in a Fortran 77 Program with arrays using the Lahey Compiler and Linker

2. Fortran 77 Program Need help

3. converting old program to standard fortran 77

4. Need help in Alpha-VMS 6.2, F-77

5. Semi-OT: help ID old Fortran 77 text book

6. VAX VMS Fortran -> Fortran 77 convertor

7. Need help udating OLD fortran programs

8. Need Help: compiling old Fortran program

9. Job - Denver, CO - Fortran 77 on DEC VMS

10. Oh I miss my VMS fortran 77

11. VMS Fortran 77 Problem

12. VAX/VMS Fortran 77 extensions

 

 
Powered by phpBB® Forum Software