Closed jsevo closed 1 year ago
Fixed by:
(setq gptel-use-curl nil)
Hi @jsevo, I'm reopening this as this is a bug, avoiding curl doesn't fix it. Could you explain the problem exactly? Do you get this error:
Symbol’s function definition is void: gptel-curl-get-response
AND the status stuck on "Waiting"?
May I ask how you installed gptel? gptel-curl-get-response
should be available with a regular package-install
.
Not OP, just speaking for myself.
I had this issue too. Fixed by (require 'gptel-curl)
.
I didn't install gptel from a package archive, but simply cloned the repo and added it to load-path
, and thus the loaddefs file is not generated at all.
(I believe this is not a bug per se, but it may be helpful to add an autoload
call in gptel.el
.)
I had this same issue, also as a result of doing a manual install. (I don't want to add the melpa archive since I'm only using melpa-stable, so I didn't install through package-install.)
I had this same issue, also as a result of doing a manual install. (I don't want to add the melpa archive since I'm only using melpa-stable, so I didn't install through package-install.)
The correct way to install it manually would be to run package-install-file on the repo directory. Is there a reason you don't want to do it this way?
The only reason was I didn't know that was the correct way. Thanks for the pointer!
Sorry for the late reply. I just found your note here after encountering the same issue again.
Cool, I've updated the readme with better instructions. Thanks for bringing it up.
I got the same issue:
Querying ChatGPT...
gptel-send: Symbol’s function definition is void: gptel-curl-get-response
I can fix it be using:
(require 'gptel-curl)
(use-package gptel
:config
(require 'gptel-curl)
;; ...
)
@willbush How are you installing gptel?
@willbush How are you installing gptel?
I use https://github.com/nix-community/emacs-overlay in NixOS.
I bring that into my entire system configuration here. I define what external packages I need here. Then, I use use-package to configure it here.
Basically, I don't use a run-time package manager to fetch packages in Emacs. The command I use to update my system also updates Emacs and packages based on a flake.lock
file.
So in my load-path
I have:
"/nix/store/b94k756s95c21h5k3rvn30n8q3br6rhj-emacs-packages-deps/share/emacs/site-lisp/elpa/gptel-20231210.334"
Package gptel is external.
Status: External in ‘/nix/store/b94k756s95c21h5k3rvn30n8q3br6rhj-emacs-packages-deps/share/emacs/site-lisp/elpa/gptel-20231210.334/’ (unsigned).
Version: 20231210.334
Commit: 1f3b911c4e7c718ae5940f33ab12910e1cc7fac8
Summary: A simple multi-LLM client
Requires: emacs-27.1, transient-0.4.0
Website: https://github.com/karthink/gptel
Keywords: convenience
I'm not familiar with nix's method of handling emacs packages. Does it generate autoloads for them? It looks like you're missing some autoloads from gptel-curl.el.
Message ID: @.***>
It does. I'm not too sure what's going wrong. Perhaps you might notice something?
dired of gptel-20231210.334:
/nix/store/b94k756s95c21h5k3rvn30n8q3br6rhj-emacs-packages-deps/share/emacs/site-lisp/elpa/gptel-20231210.334:
dr-xr-xr-x 2 root root 4.0K Dec 31 1969 .
dr-xr-xr-x 135 root root 12K Dec 31 1969 ..
lrwxrwxrwx 2 root root 123 Dec 31 1969 gptel.el -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel.el
lrwxrwxrwx 2 root root 124 Dec 31 1969 gptel.elc -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel.elc
lrwxrwxrwx 2 root root 133 Dec 31 1969 gptel-autoloads.el -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-autoloads.el
lrwxrwxrwx 2 root root 128 Dec 31 1969 gptel-curl.el -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-curl.el
lrwxrwxrwx 2 root root 129 Dec 31 1969 gptel-curl.elc -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-curl.elc
lrwxrwxrwx 2 root root 130 Dec 31 1969 gptel-ollama.el -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-ollama.el
lrwxrwxrwx 2 root root 131 Dec 31 1969 gptel-ollama.elc -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-ollama.elc
lrwxrwxrwx 2 root root 130 Dec 31 1969 gptel-openai.el -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-openai.el
lrwxrwxrwx 2 root root 131 Dec 31 1969 gptel-openai.elc -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-openai.elc
lrwxrwxrwx 2 root root 127 Dec 31 1969 gptel-pkg.el -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-pkg.el
lrwxrwxrwx 2 root root 133 Dec 31 1969 gptel-transient.el -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-transient.el
lrwxrwxrwx 2 root root 134 Dec 31 1969 gptel-transient.elc -> /nix/store/0jpkpc67k0xzjk2dqbpr2jb56jrjs06b-emacs-gptel-20231210.334/share/emacs/site-lisp/elpa/gptel-20231210.334/gptel-transient.elc
gptel-autoloads.el
:
;;; gptel-autoloads.el --- automatically extracted autoloads (do not edit) -*- lexical-binding: t -*-
;; Generated by the `loaddefs-generate' function.
;; This file is part of GNU Emacs.
;;; Code:
(add-to-list 'load-path (or (and load-file-name (directory-file-name (file-name-directory load-file-name))) (car load-path)))
;;; Generated autoloads from gptel.el
(autoload 'gptel-mode "gptel" "\
Minor mode for interacting with ChatGPT.
This is a minor mode. If called interactively, toggle the `GPTel
mode' mode. If the prefix argument is positive, enable the mode,
and if it is zero or negative, disable the mode.
If called from Lisp, toggle the mode if ARG is `toggle'. Enable
the mode if ARG is nil, omitted, or is a positive number.
Disable the mode if ARG is a negative number.
To check whether the minor mode is enabled in the current buffer,
evaluate `gptel-mode'.
The mode's hook is called both when the mode is enabled and when
it is disabled.
(fn &optional ARG)" t)
(autoload 'gptel-send "gptel" "\
Submit this prompt to ChatGPT.
With prefix arg ARG activate a transient menu with more options
instead.
(fn &optional ARG)" t)
(autoload 'gptel "gptel" "\
Switch to or start ChatGPT session with NAME.
With a prefix arg, query for a (new) session name.
Ask for API-KEY if `gptel-api-key' is unset.
If region is active, use it as the INITIAL prompt. Returns the
buffer created or switched to.
(fn NAME &optional _ INITIAL)" t)
(register-definition-prefixes "gptel" '("gptel-"))
;;; Generated autoloads from gptel-curl.el
(autoload 'gptel-curl-get-response "gptel-curl" "\
Retrieve response to prompt in INFO.
INFO is a plist with the following keys:
- :prompt (the prompt being sent)
- :buffer (the gptel buffer)
- :position (marker at which to insert the response).
Call CALLBACK with the response and INFO afterwards. If omitted
the response is inserted into the current buffer after point.
(fn INFO &optional CALLBACK)")
(register-definition-prefixes "gptel-curl" '("gptel-"))
;;; Generated autoloads from gptel-ollama.el
(autoload 'gptel-make-ollama "gptel-ollama" "\
Register an Ollama backend for gptel with NAME.
Keyword arguments:
HOST is where Ollama runs (with port), typically localhost:11434
MODELS is a list of available model names.
STREAM is a boolean to toggle streaming responses, defaults to
false.
PROTOCOL (optional) specifies the protocol, http by default.
ENDPOINT (optional) is the API endpoint for completions, defaults to
\"/api/generate\".
HEADER (optional) is for additional headers to send with each
request. It should be an alist or a function that retuns an
alist, like:
((\"Content-Type\" . \"application/json\"))
KEY (optional) is a variable whose value is the API key, or
function that returns the key. This is typically not required for
local models like Ollama.
Example:
-------
(gptel-make-ollama
\"Ollama\"
:host \"localhost:11434\"
:models \\='(\"mistral:latest\")
:stream t)
(fn NAME &key HOST HEADER KEY MODELS STREAM (PROTOCOL \"http\") (ENDPOINT \"/api/generate\"))")
(register-definition-prefixes "gptel-ollama" '("gptel--ollama-context"))
;;; Generated autoloads from gptel-openai.el
(autoload 'gptel-make-openai "gptel-openai" "\
Register a ChatGPT backend for gptel with NAME.
Keyword arguments:
HOST (optional) is the API host, typically \"api.openai.com\".
MODELS is a list of available model names.
STREAM is a boolean to toggle streaming responses, defaults to
false.
PROTOCOL (optional) specifies the protocol, https by default.
ENDPOINT (optional) is the API endpoint for completions, defaults to
\"/v1/chat/completions\".
HEADER (optional) is for additional headers to send with each
request. It should be an alist or a function that retuns an
alist, like:
((\"Content-Type\" . \"application/json\"))
KEY (optional) is a variable whose value is the API key, or
function that returns the key.
(fn NAME &key HEADER MODELS STREAM (KEY \\='gptel-api-key) (HOST \"api.openai.com\") (PROTOCOL \"https\") (ENDPOINT \"/v1/chat/completions\"))")
(autoload 'gptel-make-azure "gptel-openai" "\
Register an Azure backend for gptel with NAME.
Keyword arguments:
HOST is the API host.
MODELS is a list of available model names.
STREAM is a boolean to toggle streaming responses, defaults to
false.
PROTOCOL (optional) specifies the protocol, https by default.
ENDPOINT is the API endpoint for completions.
HEADER (optional) is for additional headers to send with each
request. It should be an alist or a function that retuns an
alist, like:
((\"Content-Type\" . \"application/json\"))
KEY (optional) is a variable whose value is the API key, or
function that returns the key.
Example:
-------
(gptel-make-azure
\"Azure-1\"
:protocol \"https\"
:host \"RESOURCE_NAME.openai.azure.com\"
:endpoint
\"/openai/deployments/DEPLOYMENT_NAME/completions?api-version=2023-05-15\"
:stream t
:models \\='(\"gpt-3.5-turbo\" \"gpt-4\"))
(fn NAME &key HOST (PROTOCOL \"https\") (HEADER (lambda nil \\=`((\"api-key\" \\=\\, (gptel--get-api-key))))) (KEY \\='gptel-api-key) MODELS STREAM ENDPOINT)")
(defalias 'gptel-make-gpt4all 'gptel-make-openai "\
Register a GPT4All backend for gptel with NAME.
Keyword arguments:
HOST is where GPT4All runs (with port), typically localhost:8491
MODELS is a list of available model names.
STREAM is a boolean to toggle streaming responses, defaults to
false.
PROTOCOL specifies the protocol, https by default.
ENDPOINT (optional) is the API endpoint for completions, defaults to
\"/api/v1/completions\"
HEADER (optional) is for additional headers to send with each
request. It should be an alist or a function that retuns an
alist, like:
((\"Content-Type\" . \"application/json\"))
KEY (optional) is a variable whose value is the API key, or
function that returns the key. This is typically not required for
local models like GPT4All.
Example:
-------
(gptel-make-gpt4all
\"GPT4All\"
:protocol \"http\"
:host \"localhost:4891\"
:models \\='(\"mistral-7b-openorca.Q4_0.gguf\"))")
;;; Generated autoloads from gptel-transient.el
(autoload 'gptel-menu "gptel-transient" nil t)
(register-definition-prefixes "gptel-transient" '("gptel-"))
;;; End of scraped data
(provide 'gptel-autoloads)
;; Local Variables:
;; version-control: never
;; no-byte-compile: t
;; no-update-autoloads: t
;; no-native-compile: t
;; coding: utf-8-emacs-unix
;; End:
;;; gptel-autoloads.el ends here
I did check that gptel-curl.el
has
;;TODO: The :transformer argument here is an alternate implementation of
;;`gptel-response-filter-functions'. The two need to be unified.
;;;###autoload
(defun gptel-curl-get-response (info &optional callback)
;;; elided ...
Hmm I tried with emacs -Q
and same thing. I am going to try no native comp and perhaps installing manually.
Seems to remain stuck, with a "Waiting" status