haskell-doc-mode is not asynchronous wrt ghci process, unlike eldoc-mode

[ikirill]

haskell-mode's support for eldoc-mode is asynchronous, so placing point on top of an expression while a ghci computation is already running does not hang emacs. On the contrary, haskell-doc-mode runs the :type command synchronously, possibly hanging emacs when point is moved until the corresponding ghci command is executed.

https://github.com/haskell/haskell-mode/blob/master/haskell-doc.el#L1570

https://github.com/haskell/haskell-mode/blob/master/haskell-doc.el#L1517

[gracjan]

Good catch. There were recent changes that solved a similar issue in #607.

[ikirill]

One thing about it is that haskell-doc-mode is quite old, IIUC, and seems to be duplicating what the standard eldoc-mode is doing anyway. Since eldoc-mode is included with emacs, would it be reasonable to deprecate haskell-doc-mode and move to only using eldoc-mode? In my configuration I turned it off when it started giving me this problem.

[gracjan]

@ikirill: Can you research this topic?

If there is a standard way to provide a functionality then it is usually better to use it. On the other hand we would like to avoid upsetting our users by breaking something they liked.

[gracjan]

We will not be fixing haskell-doc-mode, we will work on eldoc-mode support instead.