`pop-to-buffer` vs `switch-to-buffer` about conda.el

[dwa]

Hi @wtianyi.
Thanks, I believe you might be right. The problem I had with switch-to-buffer was that after opening a file, I had to jump to another buffer and then back again for the conda-env-autoactivate-mode to do its job.
But now that you brought it up, I did some testing and noticed that conda-current-name gets the wrong value (with pop-to-buffer) when jumping between buffers that use different conda envs. This seem to work better with switch-to-buffer.

But what is the right solution to activating conda on the opening of a file?

As a workaround, I added conda-env-activate-for-buffer to find-file-hook:

(use-package conda
  :hook
  (find-file . conda-env-activate-for-buffer)
  :config
  (conda-env-initialize-eshell)
  (conda-env-autoactivate-mode t)
  :ensure t)

So maybe that commit should be reverted, neither solution seem perfect, but at least the problem with switch-to-buffer can be worked around.

[wtianyi]

Thanks for your detailed explanation and tests! @dwa

I do resonate with you on the problem that the environment is not auto-activated on the opening of a file, and I too don't have a satisfying solution.

I was thinking that maybe there is some way to make "conda.el" work with "projectile" so that the conda environment is consistent and set "per project". But I have not nearly enough knowledge about "projectile" to make any proposals.

Thanks for your workaround with find-file-hook too!

[wtianyi]

I'm closing the issue. Should I have a better idea about projectile I'll potentially open a new issue or pull request.

[wtianyi]

Referring related open issues #29 #64 #81