haskell-indentation-mode does not bind DEL

[drvink]

Sorry for the noise, but I've looked over the commit log and I can't tell if I'm losing my mind or if something's changed with the default behavior of DEL under haskell-indentation-mode. It doesn't seem to be getting rebound for me anymore (when I C-h k DEL, it's bound to delete-backward-char), and furthermore, the cursor gets placed differently from where I expect it used to be.

The info pages seem to indicate that DEL should be getting rebound:

‘haskell-indentation-mode’

     Improved variation of ‘haskell-indent-mode’ indentation mode.
     Rebinds <RET> and <DEL>, so that indentations can be set and
     deleted as if they were real tabs.

Looking through the history, the behavior I'm used to from DEL was supplied by haskell-indentation-delete-backward-char, but the only relevant commit I see is c91c6b2, which was in March.

I update haskell-mode from master pretty often, so I don't know why I'm only seeing this now, but it's driving me crazy. Can we please have the old stuff back as an alternative, or can you tell me what functions I can use (if any) that will give similar behavior to the way DEL and the indentation algorithm for a new line used to work?

[drvink]

The stuff controlled by this was what I was referring to:

(defcustom haskell-indentation-delete-backward-jump-line nil
  "Delete backward jumps to the previous line when removing last indentation."
  :type 'boolean
  :group 'haskell-indentation)

(defcustom haskell-indentation-delete-indentation t
  "Delete removes indentation."
  :type 'boolean
  :group 'haskell-indentation)

...and I preferred the previous indent placement algorithm by far.

[gracjan]

Hi, thanks for the report.

S-TAB is bound to haskell-indentation-indent-backwards, is that the functionality you are looking for?

If you have specific case where indentation is subpar please report it! Current code is hackable and has a test suite so that gives good prospects for future of haskell-indentation.

[drvink]

haskell-indentation-indent-line and haskell-indentation-indent-backwards achieve part of it, but they're not quite the same as what haskell-indentation-delete-backward-indentation, haskell-indentation-delete-indentation, or haskell-indentation-delete-backward-jump-line did, since they only handle indent and don't combine indenting with deletion. I need to test more for the indent placement problem--I can't reproduce the strange behavior I was previously having, so I may have been mistaken.

I updated the old deletion code to work on current haskell-mode.

[gracjan]

Is this the same request as in #866? If so I'd like to understand how this is supposed really work, I've tried to use it and it was horrible.

[drvink]

That was part of it, but not all, if I recall. I filed this issue a while ago. :)

[gracjan]

The consensus so far is that binding DEL is not the best usability option and should not be done. S-TAB works more predictably.