Why? you may ask.
Well, my current project (i’m semi-retired) involves a bunch of machines talking to each other, using a framework built on PetriNet , which is grand fun. And some of those machines are cheap.
So out of curiosity I wondered if my dev environment would work on those.
The answer is Yes; on the pi w0, which is arm6, cheap and very competent for the price.
Not that you’d use it for dev if you could avoid it; I just wanted the familiar key sequences wherever.
I had to compile neovim; it took an age of course and I had to grab a couple of things like libtool and gettext. Compiling the correct version of node for CoC was not on - anyway waiting for CoC to do anything would be a paint drying experience.
But treesitter works, pop-up windows, lots of good stuff. Slowish, but quite usable - and this is V0.6 with debug on.
I’m impressed, with linux itself, all the tools around it, and with the neovim project
1 Like
Cool!! Happy it works. If you want some kind of dev environment then you are better of using the built-in LSP client, instead of Coc in this specific resource constraint environment. Since the latter required nodejs and that has a lot of overhead in all aspects for the kind of machine you are working on. In contrast the built in is written in Lua which is bundle with nvim.
Now on the language server side of things you could run them in the machine if they are light, or you can run them in another machine and send the results over the network if they are too heavy.
Have fun. For more info about the built in language server you can read :help lsp
I’ll get to the inbuilt lsp soon. At the moment, all heavy lifting I do on reasonably sensible machines, so all can be cross-compiled when needed. I just did this out of stubborness. It’s done in java, C and rust, so all I need to program and test on the wee small is the last step to the metal.
1 Like