bfredl

hmm weird, and always moves to the _first_ line, even if you do it many times? Just for debugging try add `return` as the first line in `append_outbuf` in `rplugin/python/nvim_ipy.py`...

Did you try with disabled `append_outbuf` as I suggested above?

If you just remove this part ``` for w in self.vim.windows: if w.buffer == self.buf: w.cursor = [len(self.buf), int(1e9)] ``` does it still happen?

Unfortunately `step2.setStdIn(.{ .lazy_path = foo_header })` does not work as a workaround. it still adds the full generated path of `foo_header` to the hash, same as for `addFileArg(foo_header)`

Nice, I took the liberty to update the issue header with some descriptions.

It is somewhat a hack, it uses time addressed undo history. Ideally `nvim-miniyank` shouldn't clutter undo history at all. This functionally exists internally in nvim, `u_undo_and_forget`, but it is not...

Acually, binary plugin could just be lua .so module. Only thing we need is a lua helper function "call this callback when this fd is ready for reading", ideally at...

I implemented binary plugin concept here: https://github.com/bfredl/nvim-x11 . It is a hack though: it steals the address of the main uv loop through the symbol `main_loop`. Also lua cannot directly...

How does xsel not bring X11 into the picture? > We've come so far with the jobs-based approach, and its flexibility has been useful. I can't remember that I ever...

And BTW I'm not sure of the distribution model, but distros can safely package `nvimX11.so` in the main `neovim` package. If libXt is missing the `ffi` load fails, and we...