roxma / nvim-completion-manager

:warning: PLEASE USE https://github.com/ncm2/ncm2 INSTEAD
MIT License
917 stars 49 forks source link

Switch to multi-thread implementation for memory efficiency? #53

Closed roxma closed 7 years ago

roxma commented 7 years ago

NCM chooses multi-process implementation at first for implementation simplicity.

But multi-thread should be more memory efficient.

Since the cpython implementation has GIL, it would still be nice and more efficient to allow standalone process for some sources with heavy calculation in python, eg, jedi library.

roxma commented 7 years ago

Done.