daedsidog

> Its scope is global in Emacs, but can be set buffer-locally if required. It's difficult for me to think about local context buffers, since (right now) the process of...

> I don't follow. A truncated list of buffer names in the menu is better than no indication that additional context is active. I understand what you meant now. >...

Thanks. That answers all the unknowns I had. I'll work towards that.

> I've created a feature-context branch in this repo, please switch the PR destination to this branch? I hope that's possible to do from the Github UI without having to...

@karthink I've reformed the context adding & sampling. ~~Would you be willing to add some sort of skeleton to the transient UI that I can work from? I don't want...

@karthink This PR is nearly complete. You can ignore my previous message as I already took the liberty of adding a transient menu of my own flavor. I've done most...

@karthink I think the current state is satisfactory. * When choosing to inject context into user prompts, it will always be injected into the last user message. * I added...

> We need a better approach for attaching the context string to user messages -- right now you do it in gptel--parse-buffer in gptel-openai. Under this design it will have...

Could perhaps benefit from the features I have on my PR, i.e. removing context, highlighting context, dedicated context buffer, etc? I think those don't really conflict with the vision of...

#256 is close to fruition and I think it adds everything you want, making this PR obsolete in my opinion.