Type Enclosing Command

[xvw]

The branch exposes the ocaml-eglot-type-enclosing function, which displays the type of the expression under the cursor.

• ocaml-eglot-type-enclosing-increase-verbosity (C-c
  C-t): to increase the verbosity of the type observed
• ocaml-eglot-type-enclosing-grow (C-↑): to grow the
  expression
• ocaml-eglot-type-enclosing-shrink (C-↓): to shrink the
  expression
• ocaml-eglot-type-enclosing-copy (C-w): to copy the
  type expression to the kill-ring (clipboard)

Decisions

At present, the control of the number of fines is reset to zero each time the enclosures are varied. However calling C-c C-t will increase the verbosity on the current enclosing type. It's easily possible to maintain verbosity control by ‘requests’ if you think that's better (which I don't).

More

Currently, there are no queries to ask for the type of an identifier (initially usable with the prefix argument in Merlin, ie: C-u C-c C-t will ask for an identifier, List.map and will display ('a -> “b) -> ”a list -> ’b list.

But as it's possible to search by type, I'm not sure that the feature is ‘so necessary’.

[gr-im]

At present, the control of the number of fines is reset to zero each time the enclosures are varied

WDYT about keeping verbosity from enclosings to enclosings? And restoring it to 0 only when re-triggering the command? (and offering a transient binding for C-<left> and C-<right> to increase/decrease verbosity for current enclosing?)

[voodoos]

I had similar thoughts when integrating the feature in vscode and I don't have a strong preference:

• The current behavior is the same as the existing client (I think), so memory muscle will be rewarded.
• The proposed 2D navigation scheme with the arrows and verbosity memory does seems nicer.

I mean, it's probably the best time to introduce this kind of UX changes, as long as we keep the "multiple presses on C-t augments verbosity"...

[xvw]

@gr-im, @voodoos
Behaviour appended on 48c95e0

[voodoos]

Just some doc comments right now, I will try to do a more in depth-review tomorrow.

We should really pay attention to the docstring of interactive functions :-)

[voodoos]

Thanks, looks good to me. I have one last question about the handling of verbosity...

[gr-im]

Thank you so much for considering my suggestion for the verbosity, it really will be more ergonomic!
In an ideal world, it would be nice to limit it when it calculates a type identical to the previous one (so we don't have to go back too far) but I think that can be left to another PR?