search

PDP-10 / its

Incompatible Timesharing System
Other
858 stars 81 forks source link

Added build of Zork (616 point version + 100 point endgame) that uses MDL 55 #2150

Closed eswenson1 closed 1 year ago

larsbrinkhoff commented 1 year ago

Fantastic! A long-time dream coming true.

Good idea putting the zork build script in its own file.

larsbrinkhoff commented 1 year ago

CC @atsampson, @heasm66, @ZoBoRf

larsbrinkhoff commented 1 year ago

I would like to suggest to rearrange things a little bit. For example, there's a Y2K fix that's logically separate from the rest. But I know it's painful to rewrite git history, so I can do it myself if that's ok?

eswenson1 commented 1 year ago

Sure! Go for it!

larsbrinkhoff commented 1 year ago

Eric, is this XXFILE useful here? https://github.com/PDP-10/its/issues/832#issuecomment-688485007

eswenson1 commented 1 year ago

I wasn't able to get that XXFILE to work. First of all, there is no ITS version of LITLPK available (that I can find). The version that I can find is for TOPS-20 (it references ). And since this mini package manager only looks in ONE directory for an FBIN, and the XXFILE does a <USE "CLEAN" "PURITY"> and nowhere can I find evidence that those two packages lived in the SAME (one) directory, that can never be satisfied by the implementation in LITLPK.

RRS has hacked a version of LITLPK -- but his, too, only looks in one directory and only for ">" files, and not FBINs or NBINs, it isn't immediately useful as a substitute.

Finally, this XXFILE uses a saved image (MUDDLE;M55UNI SAVE), which I can find in ToTS, but for which I can't find a means to recreate.

The XXFILE that I conjured up works -- albeit without all the optimizations that CFS;MADADV XXFILE has. However, with our ITS systems' larger memory (than DM), they appear not to be necessary. Also, we're not as concerned about sharing pure images. I haven't been able to get GLUE and GBINs to work yet -- another reason why this XXFILE is not usable.

larsbrinkhoff commented 1 year ago

I think it's clear that historically, MADADV SAVE was first in MADMAN and later in CFS. The switch-over was around February 1979. I think this would allow us to put the older 1977-78 binaries in MADMAN, and the freshly built save file in CFS. So the new file name wouldn't collide with the old.

Checking the records, I see that NBIN files were a little all over the place. Early on in a user directory or MUDBUG, then in LCF, and later in CFS.

This information is from the .TAPE files, which has more data than the few tape images we have.

larsbrinkhoff commented 1 year ago

The last "real" TS ZORK from January 1979 loads MADMAN; MADADV SAVE. It doesn't have any assembly info. There's a newer TS RBYE (previously discussed: https://github.com/PDP-10/its/issues/832#issuecomment-457338954) which loads CFS; MADADV SAVE. It was assembled by MARC from AR1: CFS; ZORK 2.

So we have a somewhat complicated situation.

larsbrinkhoff commented 1 year ago

Just a tiny rewrite, all done now. Running yet another build locally to make sure Zork still runs fine.

eswenson1 commented 1 year ago

Thanks Lars. I was still waiting because “some checks haven’t completed yet”. Do we care?

larsbrinkhoff commented 1 year ago

They were completed now. Sometimes you have to reload the page to see the updated check statuses.

eswenson1 commented 1 year ago

Ah, ok. Thanks for merging.

-- Eric (KC6EJS)

From: Lars Brinkhoff @.> Reply-To: PDP-10/its @.> Date: Wednesday, February 22, 2023 at 10:33 PM To: PDP-10/its @.> Cc: Eric Swenson @.>, Author @.***> Subject: Re: [PDP-10/its] Added build of Zork (616 point version + 100 point endgame) that uses MDL 55 (PR #2150)

They were completed now. Sometimes you have to reload the page to see the updated check statuses.

— Reply to this email directly, view it on GitHubhttps://github.com/PDP-10/its/pull/2150#issuecomment-1441270513, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAGT4Y4IWGWBYGEKM7H3QOLWY4ADDANCNFSM6AAAAAAVCOCJOM. You are receiving this because you authored the thread.Message ID: @.***>

larsbrinkhoff commented 1 year ago

@heasm66, I don't remember exactly but you would probably know: is there any Zork source code file set in the https://github.com/MITDDC/ repositories that seems to be close enough to compile and run earlier versions? E.g. the April 1978 GSB files seem promising.

heasm66 commented 1 year ago

My main goal was to make them work under Confusion. For the 1977-12 and 1978-01, I had to patch a lot with later files and change the syntax a bit because Confusion isn't fully implemented MDL.

1978-04 Is added a bit later after I worked with Zork, but glancing at it now it seems to be a pretty complete set of files. Getting it to work under Confusion would be a little work but now that there's a toolchain for MDL55, maybe it's worth a try?

I tried documenting where I picked and choosed files from when I patched the different versions at https://github.com/heasm66/mdlzork, but that was a while ago...

larsbrinkhoff commented 1 year ago

Thanks, @heasm66. If they work under Confusion, I believe they should work (or be made to work) under MDL 55.

Nothing to jump on now, but could be a nice project for a rainy day.