Lexical autocomplete
dead10ck opened this issue · comments
Describe your feature request
According to comments in #901, autocomplete has no native implementation and relies solely on LSP. A consequence of this is that when the user has no LSP installed, or one doesn't exist for the file type you are editing, there is no autocomplete at all.
It would be nice if there was a built-in autocomplete that is based purely on the tokens in the current document, without any language awareness.
I would find this useful too, for writing text / LaTeX documents!
Could anybody please provide the latest status on this FR ?
Could anybody please provide the latest status on this FR ?
"It is not done"
I've made a simple proof-of-concept LSP that performs autocompletion for the words found in the current buffer. You can find it here https://github.com/metafates/buffer-language-server
@metafates nice! Take a look to my solution too https://github.com/estin/simple-completion-language-server
This is a very important missing feature.
I found this https://github.com/estin/simple-completion-language-server.
It use multi language server to completion.
In addition to what was said in #3328 (comment), I think if there were someone interested in implementing this, it might be a good idea to register it as a hook for the event system which should relatively compartmentalize the code from the rest of Helix.
This is the one feature (or lack of it) that is really bugging me about Helix, currently, since I'm not a big fan of installing separate "binaries" like language servers besides my text editor just to get basic functionality like word-based autocomplete. Otherwise it has been a very nice experience for a Vim user, and not having to for example disable row numbers just to be able to copy text from a terminal window over an SSH connection has been very pleasant.