Startup problem with 10/23/204 release: tries to run INTERPRET.REM.CM again

[rmkaplan]

Describe the bug
At startup, the full.sysout runs the usual greeting (system and user), but then fails because it can't find MEDLEY-DIR.LCOM. It appears to be running INTERPRET.REM.CM, which is trying to run the LDEINIT file.

To Reproduce
Steps to reproduce the behavior:
Start up the sysout in the usual way

Expected behavior
Greeting completes successfully

Context (please complete the following information):

• OS: Mac
• OS Sonoma 14.6.1
• M2 chip
• Maiko version: [e.g. commit ID from git log | head]
• IL:MAKESYSDATE: [ 23-Oct-2024 02:24:13 ]

[nbriggs]

I can't reproduce this. I made a full.sysout on the rmk36--Tedit-fifth-round-2 branch (23 October 2024), then removed my ldeinit so that it wouldn't find it if it tried, and then started medley using

cd medley && ./run-medley   -g 1200x700 -sc 1200x700 -noscroll loadups/full.sysout

and it started and loaded the MEDLEY-DIR.LCOM and showed no errors.

[rmkaplan]

I saw this in the release sysout, not one in my branches or that I constructed.

…

On Oct 25, 2024, at 2:37 PM, Nick Briggs ***@***.***> wrote: I can't reproduce this. I made a full.sysout on the rmk36--Tedit-fifth-round-2 branch (23 October 2024), then removed my ldeinit so that it wouldn't find it if it tried, and then started medley using cd medley && ./run-medley -g 1200x700 -sc 1200x700 -noscroll loadups/full.sysout and it started and loaded the MEDLEY-DIR.LCOM and showed no errors. — Reply to this email directly, view it on GitHub <#1863 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJJ6J3G34YYCOGFHTHTZ5K23HAVCNFSM6AAAAABQQI2TTSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMZYHEYTMMRTGI>. You are receiving this because you authored the thread.

[masinter]

the initialization is complex because there are a lot of possibilities and the details matter.
The transcript from the shell when running the 'medley' script is useful to diagnose:

"/home/larry/il/medley/../maiko/linux.x86_64/lde" "/home/larry/il/medley/loadups/full.sysout" -id "default" -title "Medley Interlisp " -g 1880x1016 -sc 1880x1016 -noscroll
MEDLEYDIR: "/home/larry/il/medley"
LOGINDIR: "/home/larry/il"
GREET FILE: "/home/larry/il/medley/greetfiles/MEDLEYDIR-INIT.LCOM"
VMEM FILE: "/home/larry/il/vmem/lisp.virtualmem"
requested width: 1888, height: 1016
initialised
Window created
Creating renderer...
Creating texture...
SDL initialised
Unsupported keyboard type: 14
Configuring keyboard for type-3

in particular, what is GREETFILE (= LDEINIT) and the initial content of it ... INTERPRET.REM.CM is called when the contents of $LDEINIT starts with a " double quote, not a "(" left paren

[rmkaplan]

Did you see the same problem, starting up the release system from last Monday/Tuesday ?

…

On Oct 26, 2024, at 7:56 AM, Larry Masinter ***@***.***> wrote: the initialization is complex because there are a lot of possibilities and the details matter. The transcript from the shell when running the 'medley' script is useful to diagnose: "/home/larry/il/medley/../maiko/linux.x86_64/lde" "/home/larry/il/medley/loadups/full.sysout" -id "default" -title "Medley Interlisp " -g 1880x1016 -sc 1880x1016 -noscroll MEDLEYDIR: "/home/larry/il/medley" LOGINDIR: "/home/larry/il" GREET FILE: "/home/larry/il/medley/greetfiles/MEDLEYDIR-INIT.LCOM" VMEM FILE: "/home/larry/il/vmem/lisp.virtualmem" requested width: 1888, height: 1016 initialised Window created Creating renderer... Creating texture... SDL initialised Unsupported keyboard type: 14 Configuring keyboard for type-3 in particular, what is GREETFILE (= LDEINIT) and the initial content of it ... INTERPRET.REM.CM is called when the contents of $LDEINIT starts with a " double quote, not a "(" left paren — Reply to this email directly, view it on GitHub <#1863 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJJRX4KBJXDT7ARFWPTZ5OURTAVCNFSM6AAAAABQQI2TTSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMZZGYYDSNBYGI>. You are receiving this because you authored the thread.

[masinter]

i'm saying I didn't see the problem and can't reproduce it, starting from the latest release, and need more info to give any advice about what's going on.

[rmkaplan]

I tried it again, in the most recent release. I downloaded the loadups.tgz and runtime.tgz, merged them. Then ran

kaplan@Ron-Macbook-Air git-medley % rmedley
"/Users/kaplan/Local/medley3.5/release-medley/../maiko/darwin.aarch64/lde" "/Users/kaplan/Local/medley3.5/release-medley/loadups/full.sysout" -id "1" -title "Release Medley:: 1" -g 1920x1036 -sc 1888x1004
MEDLEYDIR: "/Users/kaplan/Local/medley3.5/release-medley"
LOGINDIR: "/Users/kaplan/lisp"
GREET FILE: "/Users/kaplan/Local/medley3.5/release-medley/greetfiles/MEDLEYDIR-INIT.LCOM"
VMEM FILE: "/Users/kaplan/lisp/vmem/lisp_1.virtualmem"

These screen shots show the beginning of the greeting and the break at the end. Below I'm also providing my rmedley script

export medleyc="/scripts/medley/medley.command $screen -i +"
alias rmedley=' /bin/sh $RMEDLEYDIR/$medleyc -t Release\ Medley%i --full '

This is the way I'm calling Frank's medleyc when I start up the Medley clone my working directory loadups

[rmkaplan]

A propos of the meeting today, I downloaded the mac universal and ran ./medley directly in the medley/ subdirectory. The presumably runs the maiko that is part of the download.

I'll note as a first observation that I did not run the xattr for maiko/darwin-universal/lde* that is specified in the How to run on Mac web page, and the system came up as expected. So it is not clear that that particular xattr is necessary. Should that instruction be removed?

I ran into some other problems with my personal Init, because this configuration of files is not located properly relative to the working directory that my init also looks for. So I have to make some other adjustments before I can verify whether or not the MEDLEYDIR-INIT is being run twice.

[rmkaplan]

Some more information, again running ./medley at the medley/ subdirectory of the univeral-mac download, with nothing else changed. No xattr's done.

Here is the screen image at the beginning of the startup. Note that it is in the middle of running the MEDLEYDIR-INIT.LCOM from the downloaded directory

I put in HELP's at the beginning and end of my personal INIT. In the HELP, I confirmed the existence of the MEDLEYDIR-INIT.LCOM file:

It ran thru to the HELP at the end (stumbling a little bit on some personal things that it can't find from the downloaded location). And then it finished my personal init...and tried again for the MEDLEYDIR-INIT.LCOM. I broke it while it was doing the exchange pups as part of a spell-correction for that unfound-file:

So:

Should it have been trying again, after it has already done the site greeting and my personal init?

If so, why cannot it not find the file the second time when it could find and execute it the first time?

[rmkaplan]

(Note BTW that MEDLEYDIR-INIT.LCOM is the name of the file in the OPENSTREAM under INTERPRET.REM.CM

[masinter]

The GREET code as currently defined will try to read the file named by LDEINIT (a unix file name but treated as {DSK}). Maybe changing LDEINIT to explicitly have a {DSK}?
The quarantined files -- are they INFILEP? Even when they give an error when you try to OPENFILE?

[rmkaplan]

MEDLEYDIR-INIT.LCOM is must be INFILEP becaise it is read to start the greeting (presumably because it is the value of LDEINIT (without {DSK}). But it is apparently not INFILEP when it breaks on the second attempt, after my personal init has completed.

…

On Dec 30, 2024, at 11:53 PM, Larry Masinter ***@***.***> wrote: The GREET code as currently defined will try to read the file named by LDEINIT (a unix file name but treated as {DSK}). Maybe changing LDEINIT to explicitly have a {DSK}? The quarantined files -- are they INFILEP? Even when they give an error when you try to OPENFILE? — Reply to this email directly, view it on GitHub <#1863 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJIPJ2JZG574YXCONZ32IJEODAVCNFSM6AAAAABQQI2TTSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNRWGIYTKMRZGA>. You are receiving this because you authored the thread.

[rmkaplan]

Have you tried this on your mac? Run ./medley in the medley/ download of the mac universal.

…

On Dec 31, 2024, at 12:10 AM, Ron Kaplan ***@***.***> wrote: MEDLEYDIR-INIT.LCOM is must be INFILEP becaise it is read to start the greeting (presumably because it is the value of LDEINIT (without {DSK}). But it is apparently not INFILEP when it breaks on the second attempt, after my personal init has completed. > On Dec 30, 2024, at 11:53 PM, Larry Masinter ***@***.***> wrote: > > > The GREET code as currently defined will try to read the file named by LDEINIT (a unix file name but treated as {DSK}). Maybe changing LDEINIT to explicitly have a {DSK}? > The quarantined files -- are they INFILEP? Even when they give an error when you try to OPENFILE? > > — > Reply to this email directly, view it on GitHub <#1863 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJIPJ2JZG574YXCONZ32IJEODAVCNFSM6AAAAABQQI2TTSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNRWGIYTKMRZGA>. > You are receiving this because you authored the thread. >

[nbriggs]

When I try it on macOS Ventura 13.7.1, unpacking into /tmp/241223/ this is what I see: