[Info-vax] clock problems with OpenVMS x86 on VirtualBox

[Bob Gezelter]

On Monday, May 8, 2023 at 1:06:34 PM UTC-4, John Dallman wrote:
> In article <kbsanj... at mid.individual.net>, bill.gu... at gmail.com
> (bill) wrote: 
> 
> > Because we often hear that there are still people running Vaxen 
> > with VMS on them and I expect moving a program from the VAX to 
> > x86-64 would be difficult if not impossible (especially if it 
> > has any MACRO in it!)
> If you have the source, the MACRO cross-compiler to x86 is available. It 
> had to be: chunks of VMS are still written in it. 
> 
> John
John,

A note of caution from past experience: MACRO-32 that ran on VAX processors may need a bit of work before it will compile on  Alpha/Itanium/x86-64 processors. Been there; done that. Some old VAX code is particularly crufty, e.g., full of PDP-11 like coding shortcuts that do not map well onto later processors. The parts of the OpenVMS base that are in MACRO-32 have now undergone three different wash cycles to remove problematic code.

That is not to say that it cannot be done, merely that a full audit of the code is required before undertaking the effort. Code already being compiled for Alpha/Itanium is a different story. Given the present tools situation, I would probably opt to do the initial version on an Alpha/Itanium host, moving to x86-64 after the kinks have been excised.

- Bob Gezelter, http://www.rlgsc.com