Open skaisler opened 2 months ago
In loops/users/CONVERSION-AIDS
the CONVERT-METHODS
function is using the "OLD-INTERLISP-FILE" readtable and using TEDIT to change (DEFINEQ ...)
into (DEFINE-METHOD-FNS ...)
expressions.
Likewise CONVERT-LOOPS-FILES
uses the "OLD-INTERLISP-FILE" when processing files. It appears the TRUCKIN file has been processed by this. This processing converts comments.
What I do not see is anything that would indicate why the converted comments aren't reading correctly!
There doesn't appear to be anything that changes the "INTERLISP" readtable to affect LOAD.
from https://github.com/masinter/homedir file LARRY there is a function GETLOOPS:
(GETLOOPS [LAMBDA NIL (* ; "Edited 19-Oct-2022 17:00 by lmm") (FILESLOAD LOADLOOPS) (LOADLOOPS) (COND ((CL:Y-OR-N-P "Load Truckin?") (CNDIR "../truckin") (FILESLOAD TRUCKINLOAD) (OR (BOUNDP 'TRUCKINVARS) (SETQ TRUCKINVRS)) (LoadTruckin))) (COND ((CL:Y-OR-N-P "Load sources?") (pushnew DIRECTORIES TruckinPlayersDirectory) (FOR X IN (REVERSE FILELST) WHEN (NOT (MEMBER X BEFOREFILES)) DO (LOAD X 'PROP) (MASTERSCOPE `(ANALYZE ANY ON ',X])
w;hich loads all of loops compiled and sources. Without any conversions or errors.
What's missing in this is the file LOOPSMS which should be loaded (FILESLOAD LOOPSMS) after the call to LOADLOOPS but before loading the sources and calling . ANALYZE.
After some experimenting to figure out why GETLOOPS was failing, I added to it:
(GETLOOPS
[LAMBDA NIL (* ; "Edited 2-Sep-2024 23:18 by mth")
(* ; "Edited 19-Oct-2022 17:00 by lmm")
(CL:UNLESS (AND (BOUNDP 'BEFOREFILES) (* ; "mth - added this to set BEFOREFILES")
BEFOREFILES)
(SETQ BEFOREFILES (COPY FILELST)))
(FILESLOAD LOADLOOPS)
(LOADLOOPS)
(COND
((CL:Y-OR-N-P "Load Truckin?")
(CNDIR "../truckin")
(FILESLOAD TRUCKINLOAD)
(OR (BOUNDP 'TRUCKINVARS)
(SETQ TRUCKINVRS))
(LoadTruckin)))
(COND
((CL:Y-OR-N-P "Load sources?")
(pushnew DIRECTORIES TruckinPlayersDirectory)
(for X in (REVERSE FILELST) when (NOT (MEMBER X BEFOREFILES))
do (LOAD X 'PROP)
(MASTERSCOPE `(ANALYZE ANY ON ',X])
Then it loaded LOOPS and Truckin correctly, but failed in the MASTERSCOPE analysis of Truckin.
I forgot have to add LOOPSMS after LOADLOOPS--LarryMasinter.netOn Sep 4, 2024, at 11:56 AM, Matt Heffron @.**> wrote: After some experimenting to figure out why GETLOOPS was failing, I added to it: (GETLOOPS [LAMBDA NIL ( ; "Edited 2-Sep-2024 23:18 by mth") ( ; "Edited 19-Oct-2022 17:00 by lmm") (CL:UNLESS (AND (BOUNDP 'BEFOREFILES) ( ; "mth - added this to set BEFOREFILES") BEFOREFILES) (SETQ BEFOREFILES (COPY FILELST))) (FILESLOAD LOADLOOPS) (LOADLOOPS) (COND ((CL:Y-OR-N-P "Load Truckin?") (CNDIR "../truckin") (FILESLOAD TRUCKINLOAD) (OR (BOUNDP 'TRUCKINVARS) (SETQ TRUCKINVRS)) (LoadTruckin))) (COND ((CL:Y-OR-N-P "Load sources?") (pushnew DIRECTORIES TruckinPlayersDirectory) (for X in (REVERSE FILELST) when (NOT (MEMBER X BEFOREFILES)) do (LOAD X 'PROP) (MASTERSCOPE `(ANALYZE ANY ON ',X]) Then it loaded LOOPS and Truckin correctly, but failed in the MASTERSCOPE analysis of Truckin. image.png (view on web)
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>
We knew that Interlisp/LOOPS did not like "#", but I am still confused why it did not like rando occurrences of ":".
I suggest we explicitly spell this out so we can document it in a technical not e ont he website.
Steve K.
On Thu, Sep 5, 2024 at 1:28 PM Larry Masinter @.***> wrote:
I forgot have to add LOOPSMS after LOADLOOPS--LarryMasinter.netOn Sep 4, 2024, at 11:56 AM, Matt Heffron @.**> wrote: After some experimenting to figure out why GETLOOPS was failing, I added to it: (GETLOOPS [LAMBDA NIL ( ; "Edited 2-Sep-2024 23:18 by mth") ( ; "Edited 19-Oct-2022 17:00 by lmm") (CL:UNLESS (AND (BOUNDP 'BEFOREFILES) ( ; "mth - added this to set BEFOREFILES") BEFOREFILES) (SETQ BEFOREFILES (COPY FILELST))) (FILESLOAD LOADLOOPS) (LOADLOOPS) (COND ((CL:Y-OR-N-P "Load Truckin?") (CNDIR "../truckin") (FILESLOAD TRUCKINLOAD) (OR (BOUNDP 'TRUCKINVARS) (SETQ TRUCKINVRS)) (LoadTruckin))) (COND ((CL:Y-OR-N-P "Load sources?") (pushnew DIRECTORIES TruckinPlayersDirectory) (for X in (REVERSE FILELST) when (NOT (MEMBER X BEFOREFILES)) do (LOAD X 'PROP) (MASTERSCOPE `(ANALYZE ANY ON ',X]) Then it loaded LOOPS and Truckin correctly, but failed in the MASTERSCOPE analysis of Truckin. image.png (view on web)
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>
— Reply to this email directly, view it on GitHub https://github.com/Interlisp/medley/issues/1819#issuecomment-2332274405, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARRPGX3OZICU4B3J5EN4KPTZVCIC3AVCNFSM6AAAAABNN57562VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMZSGI3TINBQGU . You are receiving this because you authored the thread.Message ID: @.***>
I added the (FILESLOAD LOOPSMS)
and it got further, but failed analyzing LOOPSKERNEL, in Object.MessageNotUnderstood
.
We knew that Interlisp/LOOPS did not like "#", but I am still confused why it did not like rando occurrences of ":".
I don't know what you mean by 'not like'. Is there (could you make sure there is) an issue describing any failure? If it's loops code you're using, then stay in an interlisp exec 9not xcl0.
Larry et al:
I was loading it in the Interlisp Exec, which is why it was balking.
The problem is transient. It might skip by several instances where we
have a date as "sm%:
Matt: Will you send the document to Larry that I sent you with examples. I stored it someplace so well, I can't find it right now.
Steve K.
On Mon, Sep 9, 2024 at 12:54 PM Larry Masinter @.***> wrote:
We knew that Interlisp/LOOPS did not like "#", but I am still confused why it did not like rando occurrences of ":".
I don't know what you mean by 'not like'. Is there (could you make sure there is) an issue describing any failure? If it's loops code you're using, then stay in an interlisp exec 9not xcl0.
— Reply to this email directly, view it on GitHub https://github.com/Interlisp/medley/issues/1819#issuecomment-2338611270, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARRPGX444S4Y3CVJMEWICMLZVXHDHAVCNFSM6AAAAABNN57562VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMZYGYYTCMRXGA . You are receiving this because you authored the thread.Message ID: @.***>
@skaisler
Are the LOOPS files you're loading files that have been modified in any way outside of Medley? (i.e., with Notepad, etc.)
You say it's transient, but that's based on not every occurance of "sm%:
I have modified them with Notepad++ to reformat for easier readability. But, that should not change the ability to load the source code.
steve K.
On Mon, Sep 9, 2024 at 8:23 PM Nick Briggs @.***> wrote:
@skaisler https://github.com/skaisler Are the LOOPS files you're loading files that have been modified in any way outside of Medley? (i.e., with Notepad, etc.) You say it's transient, but that's based on not every occurance of "sm%: " failing within a single file. Is it repeatable? Does the same file fail the same way each time it is loaded?
— Reply to this email directly, view it on GitHub https://github.com/Interlisp/medley/issues/1819#issuecomment-2339377055, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARRPGXZHPAOFS2F4NCW3ACDZVY3YZAVCNFSM6AAAAABNN57562VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMZZGM3TOMBVGU . You are receiving this because you were mentioned.Message ID: @.***>
@skaisler - but it can change the ability to load the source, since they're managed files. If you have an example of one of the LOOPS files that you have modified and that won't load, please attach it here.
@skaisler Or, instead of attaching one file here, you could commit the edited files to a branch and push it to GitHub. Don't make a Pull Request (or make one and mark it as Draft). This would allow us to investigate if there is some bizarre cross-file interaction that's contributing to the problem. (This is extremely improbable, but I've seen stranger things!)
NIck:
Here is the file that has problems:
And a screenshot:
[image: image.png] And, the trail through Truckin via PRINT statements:
[image: image.png]
And, here is where the problem occurs:
[image: image.png] Notice that the comment form is one of three forms sprinkled through the source code.
Steve K. On Tue, Sep 10, 2024 at 9:34 PM Matt Heffron @.***> wrote:
@skaisler https://github.com/skaisler Or, instead of attaching one file here, you could commit the edited files to a branch and push it to GitHub. Don't make a Pull Request (or make one and mark it as Draft). This would allow us to investigate if there is some bizarre cross-file interaction that's contributing to the problem. (This is extremely improbable, but I've seen stranger things!)
— Reply to this email directly, view it on GitHub https://github.com/Interlisp/medley/issues/1819#issuecomment-2342441810, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARRPGX5FWDA5QMBP7XV6IYTZV6MZXAVCNFSM6AAAAABNN57562VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNBSGQ2DCOBRGA . You are receiving this because you were mentioned.Message ID: @.***>
@skaisler The screenshots didn't go through.
Here is a word doc with the three images.
Steve K.
On Wed, Sep 11, 2024 at 9:22 AM Paolo Amoroso @.***> wrote:
@skaisler https://github.com/skaisler The screenshots didn't go through https://github.com/Interlisp/medley/issues/1819#issuecomment-2343653124.
— Reply to this email directly, view it on GitHub https://github.com/Interlisp/medley/issues/1819#issuecomment-2343658976, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARRPGX7QRJMEAGSDQIUWG63ZWA7YBAVCNFSM6AAAAABNN57562VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNBTGY2TQOJXGY . You are receiving this because you were mentioned.Message ID: @.***>
Attaching files with an e-mail reply doesn't seem to work - none of the source file, original images, or word doc got through to the github issue. Could you attach the files directly to the issue via the web interface, please.
Documented several attempts at resolving the problem here. LOOPS-LoadIssue.docx LispShots.docx TRUCKIN.txt
Expected behavior Should not balk on new comment style format, e.g.. (* : "comment including :")
This is transient.
Context (please complete the following information):
Additional context Add any other context about the problem here.