I believe this is an important question...
I have more than 30 Vim plugins, and I believe it's also true for many others.
It would be best if Neovim can be a drop-in replacement for Vim and work with (modern) plugins (and plugin managers, like Pathogen or Vundle) the same way as before.
So, I'd like to raise an issue: how to ensure that most plugins still work and they work well together (I am talking about the most important/popular plugin/scripts)? How do we detect incompatibilities as soon as possible to minimize the cost of fixing it?
Here's some idea:
Any other ideas, plans or proposals?
A first-level test is of course that the developers themselves usually have a lot of plugins installed (about 20 in my case), and in the interest of dogfooding, everybody should start using nvim asap. I already do it. Once we have a better test infrastructure, CI tests would be even better, of course.
Protip for dogfooders with MacVim installed: set VIMRUNTUME to /path/to/MacVim.app/Contents/Resources/vim/runtime
Yes, the compatibility is important.
The tests are useful for regression. But it is not perfect.
We must make effort like NFA Regexp engine feature.
I think this will mostly resolve itself when everyone can use neovim to develop neovim. I would think everyone developing neovim is using vim and have many useful plugins installed and configured.
Agreed with @simendsjo and running against time to make sure we can all use neovim soon!
@simendsjo @tarruda are you guys not using nvim? I'm using it to develop neovim since day 1. :)
@philix Great to know you are already using it. I'm not because I still depend on some python plugins. I'm hurrying to implement the python compatibility layer soon so I will be able to use it myself :)