Closed M4rtinK closed 2 years ago
Pre-release pythons are not supported.
That said, we recently upgraded bottle to the latest. Have you updated ycmd to latest and its submodules?
Duplicate https://github.com/ycm-core/YouCompleteMe/issues/4056
I just built ycmd with python 3.11.0b3 at the latest commit and it works fine.
Printing YouCompleteMe debug information...
-- Resolve completions: On demand
-- Client logfile: /var/folders/mr/p_8nf2d90gvd72mh012fqlg00000gn/T/ycm_qt__jt_w.log
-- Server Python interpreter: /Users/ben/.pyenv/versions/3.11.0b3/envs/3.11-beta-bottle/bin/python3
-- Server Python version: 3.11.0b3
-- Server has Clang support compiled in: True
-- Clang version: clang version 15.0.1 (https://github.com/tru/llvm-release-build 33b6cbead48e63164b3e7c5ac9d34505c0391552)
-- Extra configuration file found and loaded
-- Extra configuration path: /Users/ben/.vim/bundle/YouCompleteMe/third_party/ycmd/ycmd/default_ycm_extra_conf.py
-- C-family completer debug information:
-- Clangd running
-- Clangd process ID: 32037
-- Clangd executable: ['/usr/local/opt/llvm/bin/clangd', '-background-index', '-completion-style=bundled', '-pch-storage=memory', '
So the issue is you need to update ycmd to latest.
So the issue is you need to update ycmd to latest. Yep, that worked - I've updated it but apparently too long ago. :) Thanks! :)
I'm using the Fedora 37 Beta on my machine and I have noticed ycmd fails to start. When debugging the issue, I have got to this traceback:
This is likely due to the pre-release Python 3.11, where
getargspec
apparently is no longer available in inspect:Looking at the docs, 3.11 no longer lists the
getargspec
but refers to it:https://docs.python.org/3.11/library/inspect.html
3.10 docs mark it as deprecated:
https://docs.python.org/3.10/library/inspect.html#inspect.getargspec
So while this is pre-release distro & pre-release version of Python, this might still become a bigger issue once more stable distros switch to 3.11, it it indeed means this function is dropped in 3.11.
Also the most likely issue is the bundled bottle is just too old & more recent version might work fine on latest Python versions.