backend company-capf error

[MarinosK]

Whenever I e.g. write SinOsc.ar() and with the pointer inside the parentheses, I get this:

Company: backend company-capf error "Scan error: "Containing expression ends prematurely", 198, 198" with args (prefix)

Debug trace:

Debugger entered--Lisp error: (scan-error "Containing expression ends prematurely" 198 198) scan-sexps(199 -1) forward-sexp(-1 nil) backward-sexp(1) sclang-completion-at-point() completion--capf-wrapper(sclang-completion-at-point optimist) company--capf-wrapper(sclang-completion-at-point optimist) company--capf-data-real() company--capf-data() company-capf(prefix) apply(company-capf prefix) company-call-backend-raw(prefix) company--force-sync(company-call-backend-raw (prefix) company-capf) company-call-backend(prefix) company--begin-new() company--perform() company-auto-begin() company-idle-begin(#<buffer *SCLang:Workspace*> #<window 3 on *SCLang:Workspace*> 426 199) apply(company-idle-begin (#<buffer *SCLang:Workspace*> #<window 3 on *SCLang:Workspace*> 426 199)) timer-event-handler([t 26107 20089 638185 nil company-idle-begin (#<buffer *SCLang:Workspace*> #<window 3 on *SCLang:Workspace*> 426 199) nil 0 nil])

I use GNU Emacs 29.1 (build 1, aarch64-apple-darwin23.1.0, Carbon Version 170 AppKit 2487.2) of 2023-11-11

[kflak]

Can confirm the same error, using corfu and cape.

System: arch linux, emacs 29.4

[dylangleason]

Also can confirm on Linux 6.10.10-arch1-1 and using Emacs 29.4

[kflak]

Emacs 31, same story...

[buenchaval]

Can confirm with Emacs 29.4 and Gentoo Linux.

[codecollider]

I'm getting the same messages. Ubuntu 24.04 Gnome, Emacs 30.2, scel installed via option 1 according to the read.me

[alt-ee]

I've just started using scel and noticed the same issue. The problem seems to be that the sclang-completion-at-point function uses backward-sexp(1) to get the start position of the current word. This works if the point is somewhere inside a word, but not if the point is immediately after a "(". You can see this interactively in a scratch buffer by inserting "(some-text)", put your point on the "-" and run backward-sexp (C-M-b by default). The point will move to the start of the text. Run it again and you will get the message "No previous sexp", which I believe is just the interactive equivalent to "Scan error: "Containing expression ends prematurely"".

So that function is actually working as expected, the problem is that sclang-completion-at-point doesn't handle this case elegantly so the function just exits. I'm not sure what the preferred way to handle this would be. If you (ignore-errors (backward-sexp 1) t) then the error is suppressed and completion seems to run on an empty string, i.e. you are shown the full completion list of every possible sclang symbol. The help page for completion-at-point-functions states that they should return nil if no completions are found, which seems like the more "correct" option as there is nothing to complete when the point is immediately after an opening bracket.

I'm also not sure why backward-sexp is being used for this purpose rather than backward-word, as sclang doesn't use sexps and the behavior seems identical for sclang code, other than producing this error in instances where the point is immediately after a bracket? I'm pretty new to sclang though so there might be some syntax I'm not aware of where this makes sense.