Allow inserting closing delimiters for non-autoskip pairs

[johnmastro]

Don't inhibit inserting closing delimiters for pairs that don't use the autoskip action.

Commit a677859 added code to prevent closing delimiters from being inserted in strict mode. Typing a closing delimiter will either skip over the delimiter that was automatically inserted, or emit a message that an unbalanced delimiter was blocked from insertion. I think it makes sense to exempt from this pairs that don't use the autoskip action (and/or the insert action? more on that below), since in that case the only reason to type the closing delimiter is if you want to insert it.

I ran into this because, in C-like modes, I set up {} for the navigate and wrap actions, but not insert or autoskip.

It may also make sense to exempt pairs without the insert action, for basically the same reason - if you've configured Smartparens not to insert the closing delimiter for you, you probably intend to type it yourself. However, I never use insert without autoskip or vice versa, so I'm not sure. Of the two, I went with autoskip because it is more directly connected to typing a closing delimiter.

[johnmastro]

For a test, would something like this be appropriate?

(ert-deftest sp-test-sp-dont-inhibit-closing-pair-unless-autoskip ()
  (sp-test-with-temp-buffer "" nil
    (let ((sp-pairs sp-pairs))
      (smartparens-strict-mode)
      (sp-local-pair major-mode "{" "}" :actions '(insert autoskip navigate wrap))
      (execute-kbd-macro "}")
      (should (eq (char-before) nil))
      (sp-local-pair major-mode "{" "}" :actions '(:rem autoskip))
      (execute-kbd-macro "}")
      (should (eq (char-before) ?\})))))

[Fuco1]

I think we can go with autoskip for now. Later I have some thoughts of making the system a bit more fine-grained so we can have a permission for autoskip and for the insertion inhibition as separate things (as well as the regular insert action)

[Fuco1]

We can pass the keyword as an argument to sp-get-pair which I think is slightly better if we later decide to change the internal representation.

[Fuco1]

The test is fine if it works :)

[johnmastro]

Thanks for taking a look and for the feedback. I force-pushed an updated commit using the optional argument for sp-get-pair I had missed plus adding the test.

[Fuco1]

Great! Sorry it took me so long.

In the future don't hesitate to @ ping me if I take too long. My github workflow suck and I get tasks lost so often :/

[johnmastro]

No need to apologize but I will ping in the future - didn't want to be a bother if you just hadn't had a chance to get to it :) Thanks for reviewing/merging!