About the history log

[yuchanns]

          However, according to @aarnphm's suggestion, we should use [shada](https://neovim.io/doc/user/starting.html#shada-file) to store the history. I'm concerned that future refactoring might disrupt your current efforts. Are you interested in doing this refactoring?

Originally posted by @yetone in #18 (comment)

We previously discussed the history log behavior of avante.nvim.

Using shada to store history is effective, but I would propose a more flexible approach that allows users to customize their history storage. For instance, some users might prefer to save and load from an abstract storage solution with cloud integration. The default option would still be shada.

What are your thoughts?

[yuchanns]

What I mean is that we can provide a callback function to allow users to customize the save and load behavior. We do not need to implement the details of integration.

[yetone]

‌‌‌‌‌Saving in the cloud is a good feature. If the history is stored in the cloud, how can we uniquely locate a specific file within a project? Or to rephrase the question: what should the project ID be? It shouldn't be an absolute path, right?

[aarnphm]

Saving to the cloud is only useful for nvim --remote fwiw.

shada will be managed by nvim and with 0.11 dev it is pretty much free lunch as we don't have to worry too much about IO.

afaik doing anything IO intensive (such as saving history and polling from the cloud in neovim) is often frowned upon. If you know of any plugins that do this well, I'm more than happy to be proven wrong.

[aarnphm]

I was looking into this a bit, but now the logs is refactored to $XDG_DATA_STATE/avante, we can keep saving json for now.