Re: QFileSvr.400 usage

[Tom Liotta <qsrvbas@xxxxxxxxxxxx>]

Peter:

Nice discussion and I agree with your clarifications. And to tell 
the truth, I hadn't even looked at the REPLACE(*YES/*NO) before. 
When I first needed to code a CPY command, ISTR, either the parm 
didn't exist yet or there were problems with it that made it 
unreliable. Perhaps I never went back to see how CPY had improved 
over the years.

A couple more comments...

Peter Dow (ML) wrote:

> What I would expect to work is
>
>  CPY OBJ('/QSYS.LIB/ROBDELME.LIB/DELETEME.FILE')
>          TODIR('/QFileSvr.400/GDIHQ/QSYS.LIB/ROBDELME.LIB')

But [DELETEME.FILE] is treated as a subdirectory as far as its 
members are concerned. Therefore, [DELETEME.FILE/*] _might_ be more 
appropriate? The <help> for the parm gives some interesting hints 
about what happens when OBJ() is a directory.

> Which brings up the subject my buddy & I have agreed to disagree on -- 
> is QSYS.LIB in the IFS or is the IFS in QSYS.LIB?

I _think_ that that is similar to asking whether Desktop contains 
MyComputer on a Win2K system, or MyComputer contains Desktop.

AFAIK, the concepts at work behind Desktop<->MyComputer are also at 
work behind QSYS.LIB<->IFS in high-level principle. IIRC, back when 
the IFS started out, it was kind of "inside" QSYS but the various UI 
elements forced us to perceive it otherwise. As long as programmers 
cannot access below TIMI directly, we can only see what we're shown.

That is, if you materialized the QSYS 'context' and started 
following various pointers and offsets and whatnot, you could end up 
at some kind of object that would represent "the IFS". It wouldn't 
be any kind of object that was available to HLLs (nor perhaps even 
to MI?), but it would have a type/subtype and essentially "contain" 
"pointers" to the various file system "objects", one of which would 
somewhat point back to the QSYS "object/context".

(None of that is intended to be a valid technical description. It's 
mostly just verbalization of how I made sense of things for myself.)

Now, my recall might be faulty, my info might've been wrong, and my 
reference to QSYS.LIB might need to be something more like "the 
machine context" or some other esoteric term. Further, even if it 
was like that back then, it could be totally different today. But 
for me, by looking at how Desktop<->MyComputer relate, I can at 
least make a little conceptual sense of QSYS.LIB<->IFS.

Whether the mental imagery works for anyone else...? When a pointer 
leads to a pointer that leads to a pointer that refers back to the 
first, which one is the outermost?

Tom Liotta