Error in geiser-repl-import-module #2134
Comments
|
Seems to be the same issue as #1030, will see what I can do. |
|
Ok, I found where the bug comes, it is in the definition of See |
|
Thanks for looking into it!
|
|
Pierre Neidhardt <notifications@github.com> writes:
Thanks for looking into it!
* The value for emacs-25 only works with Helm. Without Helm on Emacs 26, it will fail.
Yes, so up to geiser maintainer to fix this.
* There is no fuzzy completion.
This is because the pattern is always passed to completion-table-dynamic
so all the helm match functions have no effect.
For instance, typing "g" shows an empty list, while a lot of modules
have the "g" letter.
Same here though I have here completion on geiser modules.
… —
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.*
--
Thierry
|
Wait, how is this possible? |
|
|
|
So that's another weird difference: My list is made of They are not the same element as yours and the syntax differs. |
|
Forget it, just noticed you were in the Racket REPL. I tested with Guile. I get the exact same result as yours. Still, pressing a spaces empties the list and that's a problem. Should Geiser fix it? |
|
Yes, geiser should not pass the pattern (PREFIX) to its completion function, don't know exactly where it do this in the geiser code. |
|
OK, I've updated the discussion on GitLab. |
|
Eval the two functions, and then try in each completing-read typing "b z", one completing-read returns "baz" while the other returns nothing (apart the unknow pattern of course). (defun helm-test-cands-1 ()
(all-completions "" '("foo" "bar" "baz")))
(defun helm-test-cands-2 (str)
(all-completions str '("foo" "bar" "baz")))
(completing-read "test: " (completion-table-dynamic 'helm-test-cands-2))
(completing-read "test: " (helm-test-cands-1)) |
|
Good example, thanks! |
when a function is used as collection. * helm-mode.el (helm-comp-read): Always pass empty string to helm-comp-read-get-candidates.
|
Could fix it from Helm by always passing an empty string to the function generated by |
|
Works for me, thanks! |
|
Pierre Neidhardt <notifications@github.com> writes:
Works for me, thanks!
By the way, it seems that the geiser-completion--module-list-func patch is not needed anymore.
Which patch?
Geiser still have to fix its call to `completion-table-dynamic` (no
SWTICH-BUFFER arg).
…--
Thierry
|
|
Sorry, my mistake. You are right, geiser-completion--module-list-func must be fixed.
|
|
Pierre Neidhardt <notifications@github.com> writes:
Sorry, my mistake. You are right, geiser-completion--module-list-func must be fixed.
To clarify, there is two issues:
1) As soon we enter a character in prompt we have an error saying the
current buffer is not a REPL (due to the SWITCH-BUFFER option of
completion-table-dynamic used for
emacs 26).
2) We have no enhanced completion once 1) is fixed, i.e. only prefixed
exact match.
1) need to be fixed in geiser and 2) is now fixed in Helm.
…--
Thierry
|
Expected behavior
No error.
Actual behavior (from
emacs-helm.shif possible, see note at the bottom)Error while inputting characters in geiser-repl-import-module.
Steps to reproduce (recipe)
C-c C-iBacktraces if any (
M-x toggle-debug-on-error)The trace cannot be displayed because the "display is not ready".
Here is the error value however. Truncated because it's very long.
Describe versions of Helm, Emacs, operating system, etc.
Are you using
emacs-helm.shto reproduce this bug (yes/no):No.
The text was updated successfully, but these errors were encountered: