15 Oct
2022
15 Oct
'22
4:32 p.m.
On 10/14/22 04:09, Johnny Billquist wrote:
It would be good. Of course the simplest way is to have multiple
packages ("APL" vs. "APL-FPP" or similar) but that's a bit
"dirty".
A better approach would be a way to store those constraints,
generated and stored by "@rpm configure", which the other functions
would use to select the proper files. It could be as simple as
(transparently) pointing to a different repository of package files.
I hope the APL playing today will work fine. Sounds
like all should be
good anyway. But I realized one thing below that I maybe should comment on.
It was great! Thank you (and others) for all of your help. The guy
was able to complete one of his "bucket list" items and was very happy.
Yes,
that's how I started. I have your (wonderful) RPM system
running on all of the networked RSX machines both at LSSM and at home.
When I saw APL in the list of available packages, I thought "Oh,
easy!" ...but the executable crashes due to its FPP requirement.
This is sortof a shortcoming in RPM right now, that have been brought up
by someone else as well. At the moment RPM itself have no idea about
constraints. Basically, packages might require FPP, or supervisor mode,
or whatever. But I don't have a way of tracking such requirements, or
possibly have different variants/alternatives for different hardware or
OS features. This seems like something that maybe would be good. I've basically just started from the assumption
that I only target full
fledged RSX-11M-PLUS V4 (or later), on hardware that have FPP as well as
supervisor mode.
That's great for simh, but real hardware is a lot more variable. ;)
This is probably helpful to make more clear to people,
and maybe I
should think about/fix that the system can be used also if you have some
special hardware requirements/limitations.
Another thing around RPM is that at the moment, having packages from
multiple repositories are a bit tricky, and creating packages are not
very documented as well. I don't know how many people actually use RPM,
and I don't know what issues there might be.
Well, I can tell you that I'm the primary user of it in the "orbit"
of LSSM, but all of our guys are aware of it. I tell museum visitors
about it when it seems appropriate to bring it up in conversation;
people who understand the concept are amazed that it has been recently
implemented for PDP-11s.
I'm happy when I hear people finding it useful,
but I mostly fix things
because of/when I have specific needs of my own that I need to get fixed.
That's certainly understandable. It's a really great system and it
has helped us out at LSSM tremendously, as well as me personally at home.
LSSM has a big Itanium2 VMS machine that's our always-on HECnet node.
It also provides MOP services for our DECserver terminal servers, etc.
Because the transatlantic links can sometimes be a bit slow, and there
are several RSX-11M-PLUS systems between LSSM and my place, I maintain a
local copy of the RPM repository on that system, that I copy
periodically from MIM:: and all local instances of rpm are configured to
point at that.
What I'd really love to see is similar efforts for RSTS/E and VMS.
(and maybe even RT-11, eventually) I've not done any serious DCL coding
for probably twenty years, but maybe I could take a look at that at some
point.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA