[Info-vax] 64 bit DCL ?

[Stephen Hoffman]

On 2015-01-18 20:45:50 +0000, JF Mezei said:

> On 15-01-18 15:34, David Froble wrote:
> 
>> Depending upon perspective, I'd argue that DCL is not really tied or 
>> part of VMS.
> 
> Hoff pointed out that the CLI parsing routines, used by applications to 
> get their arguments, are part of VMS run time, not part of DCL.EXE

The CLI callbacks have documented and undocumented APIs, but the 
callbacks do call through into and do use CLI-level internal code.  
That's how a CLI works, after all.

CLIs aren't publicly documented beyond what's in the Internals and Data 
Structures Manual (IDSM) and the source listings, and CLIs can also be 
somewhat difficult to debug due to the run-time environment.    For 
details of the CLI environment and requirements, see the IDSM and dig 
around for the existing discussions of and existing references to the 
(undocumented) sys$cli system service.  The sys$cli API here has some 
similarities to the exe$getspi[1] API underneath MONITOR in various 
regards, and sys$cli could certainly be fodder for a clean-up and for 
documentation, if VSI decided that the associated effort was worth it.  
But I digress.

> So it looks to me like VMS and DCL  are married at the hip.

Other than that VMS used MCR for a whole lot of its CLI operations in 
the early years.  VMS also had DEC/shell as a CLI, too.

So depending on the perspective, DCL isn't part of VMS.  But there's a 
huge pile of existing DCL code that isn't and won't be going away any 
time soon, any more than the VAX Macro32 code in the kernel is going to 
get rewritten anytime soon.

######
[1]  The MONITOR exe$getspi API was overhauled and documented and 
became available as sys$getrmi.



-- 
Pure Personal Opinion | HoffmanLabs LLC