LispWorks backend - completion stops working and raises errors after loading Trivia #513

mdbergmann · 2022-05-21T16:10:33Z

mdbergmann
May 21, 2022

Hi.

I found some issues in the LispWorks backend in Sly, but the Slime one seems to work fine.
@joaotavora what usually is the strategy to update those backends from Slime?
Is it use the Slime one and rename swank to slynk? Or is it diffing in new stuff one by one?

Thanks,
Manfred

joaotavora · 2022-05-21T16:19:49Z

joaotavora
May 21, 2022
Maintainer

Use Git in both repos in the corresponding file and find the missing commits. Then cherry-pick in the SLIME -> SLY direction. I haven't done it in a while. It usually works fine. You may also think of investigating and describing the specific issue (though I haven't had much time lately to look at this).

5 replies

mdbergmann May 21, 2022
Author

OK, I've checked. There are actually no changes to the lispworks.lisp backend.
I'll try to find a pattern for the issue I see.

mdbergmann May 22, 2022
Author

So I've narrowed it down to a package cl-mock that I'm using in my test system.
When I load this library then something (the auto completion) gets out of whack.
Probably not a problem of Sly. But not sure what.

mdbergmann May 29, 2022
Author

Apparently it seems the problematic package is trivia.

phoe Jul 1, 2022

Do the debugger/inspector work in any way? Can you inspect the slots of the type error in question? Maybe that can shed a bit of a light.

mdbergmann Jul 2, 2022
Author

Under 'more' the stack trace is actually longer.
I've expanded it:

Error (TYPE-ERROR) printing the following condition: #<TYPE-ERROR 80100FE593>
   [Condition of type TYPE-ERROR]

Restarts:
 0: [*ABORT] Return to SLY's top level.
 1: [ABORT] Quit process.

Backtrace:
 0: [error printing frame]
      Locals:
        DATUM = TYPE-ERROR
        ARGUMENTS = #<CONS <<error printing object>> 80302A55F1>
 1: [error printing frame]
      Locals:
        SYSTEM::ESTRING = TYPE-ERROR
        SYSTEM::EARGS = #<CONS <<error printing object>> 80302A55F1>
 2: [error printing frame]
      [No Locals]
 3: [error printing frame]
      Locals:
        RESULT-TYPE = SIMPLE-STRING
        SEQUENCES = #<CONS <<error printing object>> 80501E9CC1>
 4: [error printing frame]
      Locals:
        SYSTEM::RESULT-TYPE = SIMPLE-STRING
        SYSTEM::SEQUENCES = #<CONS <<error printing object>> 80501E9CC1>
 5: (SLYNK-COMPLETION::QUALIFIED-MATCHING "QL" #<The COMMON-LISP-USER package, 744/1024 internal, 0/4 external>)
      Locals:
        PATTERN = "QL"
        HOME-PACKAGE = #<The COMMON-LISP-USER package, 744/1024 internal, 0/4 external>
        FIRST-COLON = NIL
        STARTS-WITH-COLON = NIL
        TWO-COLONS = NIL
        PACKAGE-LOCAL-NICKNAMES = NIL
        PACKAGE-LOCAL-NICKNAMES-BY-PACKAGE = #<EQL Hash Table{0} 80100FA86B>
        NICKNAMES-BY-PACKAGE = #<EQL Hash Table{49} 80100FAA63>
        DBG::|internal-flet-name-for-SORTED-NICKNAMES| = #<Closure (FLET SLYNK-COMPLETION::SORTED-NICKNAMES) subfunction of SLYNK-COMPLETION::QUALIFIED-MATCHING 80501E9D69>
        DBG::COLLECT-EXTERNAL = (("QL-CDB:CONVERT-INDEX-FILE" QL-CDB:CONVERT-INDEX-FILE (0 1) 7.1875146E-4) ..)
        DBG::COLLECT-INTERNAL = NIL
        DBG::|internal-flet-name-for-COLLECT-EXTERNAL| = #<Closure (FLET SLYNK-COMPLETION::COLLECT-EXTERNAL) subfunction of SLYNK-COMPLETION::QUALIFIED-MATCHING 8020022959>
        USE-LIST = (#<The COMMON-LISP package, 0/4 internal, 978/1024 external> #<The HARLEQUIN-COMMON-LISP package, 2/8 internal, 379/512 external> #<The LISPWORKS package, 0/4 internal, 229/256 external>)
        PACKAGE = #<The TRIVIA.LEVEL2 package, 0/256 internal, 87/256 external>
        DBG::|tail-| = (#<The TRIVIA.LEVEL1 package, 0/16 internal, 15/16 external> ..)
        SORTED-NICKNAMES = ("TRIVIA" "TRIVIA.LEVEL2")
        DBG::|do-externals-symbols-package-| = #<The TRIVIA.LEVEL2 package, 0/256 internal, 87/256 external>
        S = #<SYMBOL <<error printing object>> 80800B6D59>
        DBG::G = 49
        DBG::G = #(TRIVIA.LEVEL2:ROW-MAJOR-ARRAY* 0 0 0 0 0 TRIVIA.LEVEL2:ROW-MAJOR-ARRAY 0 0 0 TRIVIA.LEVEL2:*ARITY-CHECK-BY-TEST-CALL* TRIVIA.LEVEL2:DEFOPTIMIZER 0 0 0 0 0 0 0 TRIVIA.LEVEL2:LAMBDA-LIST 0 0 0 0 0 ..)
        DBG::G = 512
        NICKNAME = "TRIVIA"
        DBG::|tail-| = ("TRIVIA.LEVEL2")
 6: (SLYNK-COMPLETION:FLEX-COMPLETIONS "ql" "common-lisp-user" :LIMIT 300)
 7: (SYSTEM::%INVOKE :INVISIBLEP T)
 8: (SYSTEM::%EVAL (SLYNK-COMPLETION:FLEX-COMPLETIONS "ql" '"common-lisp-user"))
 9: (EVAL (SLYNK-COMPLETION:FLEX-COMPLETIONS "ql" '"common-lisp-user"))
10: (SLYNK:EVAL-FOR-EMACS (SLYNK-COMPLETION:FLEX-COMPLETIONS "ql" '"common-lisp-user") "common-lisp-user" 40)
11: ((SUBFUNCTION 1 (SUBFUNCTION 1 (SUBFUNCTION 1 SLYNK::SPAWN-WORKER-THREAD))))
12: ((SUBFUNCTION (FLET SLYNK-BACKEND:CALL-WITH-DEBUGGER-HOOK) (TOP-LEVEL-FORM 50)) ..)
13: ((SUBFUNCTION 1 SLYNK::CALL-WITH-LISTENER))
14: (SLYNK::CALL-WITH-BINDINGS ((*PACKAGE* . #<The COMMON-LISP-USER package, 744/1024 internal, 0/4 external>) (*DEFAULT-PATHNAME-DEFAULTS* . #P"") (* :TRIVIA) (**) (***) (/ (:TRIVIA)) ...) ..)
15: ((SUBFUNCTION 1 (SUBFUNCTION 1 SLYNK::SPAWN-WORKER-THREAD)))
16: (MP::PROCESS-SG-FUNCTION 0 NIL NIL)
17: [error printing frame]

Not sure how to further troubleshoot.
Would it make sense to do some debug log in SLYNK-COMPLETION::QUALIFIED-MATCHING?

joaotavora · 2022-07-02T08:06:49Z

joaotavora
Jul 2, 2022
Maintainer

Before debug log, I'd just invoke the function from the repl with those arguments. You won't have the same dynamic state, but it might be sufficient to reproduce it.

…

On Sat, Jul 2, 2022, 08:51 Manfred Bergmann ***@***.***> wrote: Under 'more' the stack trace is actually longer. I've expanded it: Error (TYPE-ERROR) printing the following condition: #<TYPE-ERROR 80100FE593> [Condition of type TYPE-ERROR] Restarts: 0: [*ABORT] Return to SLY's top level. 1: [ABORT] Quit process. Backtrace: 0: [error printing frame] Locals: DATUM = TYPE-ERROR ARGUMENTS = #<CONS <<error printing object>> 80302A55F1> 1: [error printing frame] Locals: SYSTEM::ESTRING = TYPE-ERROR SYSTEM::EARGS = #<CONS <<error printing object>> 80302A55F1> 2: [error printing frame] [No Locals] 3: [error printing frame] Locals: RESULT-TYPE = SIMPLE-STRING SEQUENCES = #<CONS <<error printing object>> 80501E9CC1> 4: [error printing frame] Locals: SYSTEM::RESULT-TYPE = SIMPLE-STRING SYSTEM::SEQUENCES = #<CONS <<error printing object>> 80501E9CC1> 5: (SLYNK-COMPLETION::QUALIFIED-MATCHING "QL" #<The COMMON-LISP-USER package, 744/1024 internal, 0/4 external>) Locals: PATTERN = "QL" HOME-PACKAGE = #<The COMMON-LISP-USER package, 744/1024 internal, 0/4 external> FIRST-COLON = NIL STARTS-WITH-COLON = NIL TWO-COLONS = NIL PACKAGE-LOCAL-NICKNAMES = NIL PACKAGE-LOCAL-NICKNAMES-BY-PACKAGE = #<EQL Hash Table{0} 80100FA86B> NICKNAMES-BY-PACKAGE = #<EQL Hash Table{49} 80100FAA63> DBG::|internal-flet-name-for-SORTED-NICKNAMES| = #<Closure (FLET SLYNK-COMPLETION::SORTED-NICKNAMES) subfunction of SLYNK-COMPLETION::QUALIFIED-MATCHING 80501E9D69> DBG::COLLECT-EXTERNAL = (("QL-CDB:CONVERT-INDEX-FILE" QL-CDB:CONVERT-INDEX-FILE (0 1) 7.1875146E-4) ..) DBG::COLLECT-INTERNAL = NIL DBG::|internal-flet-name-for-COLLECT-EXTERNAL| = #<Closure (FLET SLYNK-COMPLETION::COLLECT-EXTERNAL) subfunction of SLYNK-COMPLETION::QUALIFIED-MATCHING 8020022959> USE-LIST = (#<The COMMON-LISP package, 0/4 internal, 978/1024 external> #<The HARLEQUIN-COMMON-LISP package, 2/8 internal, 379/512 external> #<The LISPWORKS package, 0/4 internal, 229/256 external>) PACKAGE = #<The TRIVIA.LEVEL2 package, 0/256 internal, 87/256 external> DBG::|tail-| = (#<The TRIVIA.LEVEL1 package, 0/16 internal, 15/16 external> ..) SORTED-NICKNAMES = ("TRIVIA" "TRIVIA.LEVEL2") DBG::|do-externals-symbols-package-| = #<The TRIVIA.LEVEL2 package, 0/256 internal, 87/256 external> S = #<SYMBOL <<error printing object>> 80800B6D59> DBG::G = 49 DBG::G = #(TRIVIA.LEVEL2:ROW-MAJOR-ARRAY* 0 0 0 0 0 TRIVIA.LEVEL2:ROW-MAJOR-ARRAY 0 0 0 TRIVIA.LEVEL2:*ARITY-CHECK-BY-TEST-CALL* TRIVIA.LEVEL2:DEFOPTIMIZER 0 0 0 0 0 0 0 TRIVIA.LEVEL2:LAMBDA-LIST 0 0 0 0 0 ..) DBG::G = 512 NICKNAME = "TRIVIA" DBG::|tail-| = ("TRIVIA.LEVEL2") 6: (SLYNK-COMPLETION:FLEX-COMPLETIONS "ql" "common-lisp-user" :LIMIT 300) 7: (SYSTEM::%INVOKE :INVISIBLEP T) 8: (SYSTEM::%EVAL (SLYNK-COMPLETION:FLEX-COMPLETIONS "ql" '"common-lisp-user")) 9: (EVAL (SLYNK-COMPLETION:FLEX-COMPLETIONS "ql" '"common-lisp-user")) 10: (SLYNK:EVAL-FOR-EMACS (SLYNK-COMPLETION:FLEX-COMPLETIONS "ql" '"common-lisp-user") "common-lisp-user" 40) 11: ((SUBFUNCTION 1 (SUBFUNCTION 1 (SUBFUNCTION 1 SLYNK::SPAWN-WORKER-THREAD)))) 12: ((SUBFUNCTION (FLET SLYNK-BACKEND:CALL-WITH-DEBUGGER-HOOK) (TOP-LEVEL-FORM 50)) ..) 13: ((SUBFUNCTION 1 SLYNK::CALL-WITH-LISTENER)) 14: (SLYNK::CALL-WITH-BINDINGS ((*PACKAGE* . #<The COMMON-LISP-USER package, 744/1024 internal, 0/4 external>) (*DEFAULT-PATHNAME-DEFAULTS* . #P"") (* :TRIVIA) (**) (***) (/ (:TRIVIA)) ...) ..) 15: ((SUBFUNCTION 1 (SUBFUNCTION 1 SLYNK::SPAWN-WORKER-THREAD))) 16: (MP::PROCESS-SG-FUNCTION 0 NIL NIL) 17: [error printing frame] Not sure how to further troubleshoot. Would it make sense to do some debug log in SLYNK-COMPLETION::QUALIFIED-MATCHING? — Reply to this email directly, view it on GitHub <#513 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAC6PQ7PU7EOO7WJ3AVUJLDVR7YGRANCNFSM5WSAVJBA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

1 reply

mdbergmann Jul 2, 2022
Author

Yeah, ok.
I can reproduce:

CL-USER> (slynk-completion::qualified-matching "foo" (find-package 'cl-user))
; Debugger entered on #<TYPE-ERROR 8010020B6B>

The local variables are similar. Seems to fail while searching for matches in trivia.level2 package.

joaotavora · 2022-07-02T08:27:20Z

joaotavora
Jul 2, 2022
Maintainer

That's nice. Then try to narrow it down a bit further. It's most likely a bug in sly, but it can also be in lispworks (in which sly would probably be patched anyway) or some very intense mischief by trivia.

…

On Sat, Jul 2, 2022, 09:23 Manfred Bergmann ***@***.***> wrote: Yeah, ok. I can reproduce: CL-USER> (slynk-completion::qualified-matching "foo" (find-package 'cl-user)) ; Debugger entered on #<TYPE-ERROR 8010020B6B> The local variables are similar. Seems to fail while searching for matches in trivia.level2 package. — Reply to this email directly, view it on GitHub <#513 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAC6PQ2WN5CAO45ILSQM573VR74APANCNFSM5WSAVJBA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

0 replies

mdbergmann · 2022-07-02T09:12:02Z

mdbergmann
Jul 2, 2022
Author

So I've narrowed it down to a section where external symbols are processed in qualified-matching:

(do-external-symbols (s package)

(symbol-name s) fails, and s is also not printable.

A further stack-trace of this is:

Backtrace:
 0: [error printing frame]
 1: [error printing frame]
 2: [error printing frame]
 3: ((SUBFUNCTION 1 (METHOD STREAM:STREAM-WRITE-STRING (SLYNK-GRAY::SLY-OUTPUT-STREAM T))))
      Locals:
        SLYNK-GRAY::START = 0
        SLYNK-GRAY::END = 5
        SLYNK-GRAY::LEN = 8000
        COUNT = 5
        SLYNK-GRAY::FREE = 5621
        STREAM = #<SLYNK-GRAY::SLY-OUTPUT-STREAM 8160283EFB>
        SLYNK-GRAY::START = 0
        SLYNK-GRAY::END = 5
        STRING = #<SIMPLE-BMP-STRING <<error printing object>> 806005BD83>
 4: ((SUBFUNCTION (FLET SLYNK-BACKEND:CALL-WITH-LOCK-HELD) (TOP-LEVEL-FORM 126)) ..)
      Locals:
        SLYNK-LISPWORKS::LOCK = #<MP:LOCK "buffer write lock" Locked once by "slynk-worker" 81603641DB>
        FUNCTION = #<Closure 1 subfunction of (METHOD STREAM:STREAM-WRITE-STRING (SLYNK-GRAY::SLY-OUTPUT-STREAM T)) 802034FFA9>
 5: [error printing frame]
 6: [error printing frame]
 7: [error printing frame]
 8: [error printing frame]
 9: [error printing frame]
10: [error printing frame]
11: [error printing frame]
12: [error printing frame]
13: (SYSTEM::SUB-FORMAT 0 17)
14: (SYSTEM::CALL-SUB-FORMAT #<SLYNK-GRAY::SLY-OUTPUT-STREAM 8160283EFB> 0 17)
15: [error printing frame]
16: (SLYNK-COMPLETION::QUALIFIED-MATCHING "QL" #<The COMMON-LISP-USER package, 474/1024 internal, 0/4 external>)

This SIMPLE-BMP-STRING looks like this:

#<SIMPLE-BMP-STRING 806005BD83>
--------------------
Dimensions: (5)
Element type: BMP-CHAR
Total size: 5
Adjustable: NIL
Fill pointer: NIL
Contents:
0: #<CHARACTER <<error printing object>> 3BBFF>
1: #\L
2: #\I
3: #\S
4: #\T

I do not know what a SIMPLE-BMP-STRING is. Probably a LispWorks string implementation?

21 replies

mdbergmann Jul 2, 2022
Author

Hmm. OK.
Can this be worked around at a central place (I mean setting the default character element type)? Because this issue seems to manifest in different areas, rpc and completion, and maybe others.

phoe Jul 2, 2022

Yes, it seems that the call to lw:set-default-character-element-type works globally. In particular, calling it in the LW rc file before loading slynk should work around the whole affair.

mdbergmann Jul 2, 2022
Author

Ah, sorry. I think you already mentioned it: during slynk setup.
Probably in the LW backend. I can give this a try.

mdbergmann Jul 2, 2022
Author

Setting it in LW startup file makes it work.
But I think this should be in SLY/SLYNK, no?

phoe Jul 2, 2022

I think it's somewhat safe to assume that everyone who uses slynk will prefer Unicode support to 8-bit character efficiency, so, yes, IMO Slynk should set that during its setup.

mdbergmann · 2022-07-02T11:44:29Z

mdbergmann
Jul 2, 2022
Author

Anyway, thank you both for your help.

0 replies

joaotavora · 2022-07-02T12:55:57Z

joaotavora
Jul 2, 2022
Maintainer

Away from my PC (at the 🏖️ hehe) but I've just skimmed this enthralling discussion, and I gather a simple patch/pr to sly's lispworks backend is in order. If so, fantastic, and I suppose the patch would apply to slime too (except that slime doesn't immediately seem to have user code triggering this...)

…

On Sat, Jul 2, 2022, 12:45 phoe ***@***.***> wrote: I think it's somewhat safe to assume that everyone who uses slynk will prefer Unicode support to 8-bit character efficiency, so, yes, IMO Slynk should set that during its setup. — Reply to this email directly, view it on GitHub <#513 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAC6PQ3NZZZNFMPJOFETO73VSATUPANCNFSM5WSAVJBA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

LispWorks backend - completion stops working and raises errors after loading Trivia #513

{{title}}

Replies: 6 comments 27 replies

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

Select a reply

LispWorks backend - completion stops working and raises errors after loading Trivia #513

mdbergmann May 21, 2022

Replies: 6 comments · 27 replies

joaotavora May 21, 2022 Maintainer

mdbergmann May 21, 2022 Author

mdbergmann May 22, 2022 Author

mdbergmann May 29, 2022 Author

phoe Jul 1, 2022

mdbergmann Jul 2, 2022 Author

joaotavora Jul 2, 2022 Maintainer

mdbergmann Jul 2, 2022 Author

joaotavora Jul 2, 2022 Maintainer

mdbergmann Jul 2, 2022 Author

mdbergmann Jul 2, 2022 Author

phoe Jul 2, 2022

mdbergmann Jul 2, 2022 Author

mdbergmann Jul 2, 2022 Author

phoe Jul 2, 2022

mdbergmann Jul 2, 2022 Author

joaotavora Jul 2, 2022 Maintainer

mdbergmann
May 21, 2022

Replies: 6 comments 27 replies

joaotavora
May 21, 2022
Maintainer

mdbergmann May 21, 2022
Author

mdbergmann May 22, 2022
Author

mdbergmann May 29, 2022
Author

mdbergmann Jul 2, 2022
Author

joaotavora
Jul 2, 2022
Maintainer

mdbergmann Jul 2, 2022
Author

joaotavora
Jul 2, 2022
Maintainer

mdbergmann
Jul 2, 2022
Author

mdbergmann Jul 2, 2022
Author

mdbergmann Jul 2, 2022
Author

mdbergmann Jul 2, 2022
Author

mdbergmann
Jul 2, 2022
Author

joaotavora
Jul 2, 2022
Maintainer