Spacemacs

[fatlazycat]

Hi, doesn't appear to work within Spacemacs.

I assume something todo with using evil-mode ?

Thanks

[mpickering]

I'm sorry but I have not used spacemacs a lot. When I did use it though, I think that I was able to use the same emacs keybindings (for coq for example).

[fatlazycat]

Sorry should have been clearer, i can invoke either refactor fn but there is no change applied to the buffer.

Will try and take a look at some point

[mpickering]

Does it work using hlint directly? (hlint --refactor) ?

[fatlazycat]

Yep thats fine, that outputs a changed file to stdout

[cocreature]

@fatlazycat Could you provide an example file that you tried, the function you called (and also how you called it) and the point your cursor was on?

I’ve never used spacemacs, but I'll try taking a look at it if you provide an example.

[fatlazycat]

Simple example, https://github.com/fatlazycat/eventjournal/blob/master/test/JournalFileTests.hs

Cursor on the d of do. In the gutter I see 'mark set' after a C-c , r

Thanks

[mpickering]

What about if you have the cursor over the o?

On Sun, Nov 29, 2015 at 2:31 PM, Graham Berks notifications@github.com wrote:

Simple example, https://github.com/fatlazycat/eventjournal/blob/master/test/JournalFileTests.hs

Cursor on the d of do. In the gutter I see 'mark set' after a C-c , r

Thanks

— Reply to this email directly or view it on GitHub https://github.com/mpickering/hlint-refactor-mode/issues/6#issuecomment-160419413 .

[fatlazycat]

No difference

On 29 November 2015 at 18:51:25, Matthew Pickering (notifications@github.com) wrote:

What about if you have the cursor over the o?

On Sun, Nov 29, 2015 at 2:31 PM, Graham Berks notifications@github.com wrote:

Simple example, https://github.com/fatlazycat/eventjournal/blob/master/test/JournalFileTests.hs

Cursor on the d of do. In the gutter I see 'mark set' after a C-c , r

Thanks

— Reply to this email directly or view it on GitHub https://github.com/mpickering/hlint-refactor-mode/issues/6#issuecomment-160419413 .

— Reply to this email directly or view it on GitHub.

[cocreature]

Alright, I've managed to install spacemacs, but sadly I can’t reproduce this. If my d,o or anywhere inside the do block it works. Also it doesn’t matter whether I’m in insert mode or in whatever the mode after you press escape is called.

Here’s a list of things that could cause problems

• Are you running the newest version of hlint-refactor-mode?
• Is hlint and refact in one of the directories listed in exec-path?
• What happens if you run hlint-refactor-refactor-at-point manually?
• Same question for hlint-refactor-refactor-buffer
• Do you have hlint options set somewhere which might exclude redundant do from being a warning? It would be strange that it works in the terminal then, but maybe it’s an option set somewhere in your emacs config?
• If you run toggle-debug-on-error and then run it again, do you get an error and a backtrace?
• Is there anything in *Messages*?
• Do you have some weird config that could cause this? I just installed spacemacs, activated the haskell layer and installed this package.

[fatlazycat]

Hmm, odd.

Updated to latest via melpa.

Both in path.

Same problem if run manually.

I don’t set any hlint options.

Nothing in messages.

Started with a clean install and still get the same error.

Perhaps OSX or the version of emacs I’m using railwaycat/emacsmacport/emacs-mac: stable emacs-24.5-z-mac-5.13

[aijony]

This issue is old, but I think I found a fix (at least for me).

I had this:

(setenv "PATH" (concat  "~/.local/bin:" (getenv "PATH")))
(add-to-list 'exec-path "~/.local/bin/")

and I changed it to:

(setenv "PATH" (concat (file-truename "~/.local/bin:") (getenv "PATH")))
(add-to-list 'exec-path "~/.local/bin/")

I guess it needed to be an absolue file path :/

[tonyday567]

I use spacemacs and hlint-refactor-refactor-buffer works most of time. Works is not really the right term - it is a thing of joy not to have to manually fix redundant brackets and bracket moving to avoid $. Except when the file seems to be too large, or there are many hlints, in which case it doesn't make any changes and the file flashes an unpleasant yellow hue. I then figure its hlint telling me to reduce complexity levels.

[georgewsinger]

I'm also having the same problem 2 years later :scream:

I have verified that hlint is installed and that hlint --refactor [filename.hs] outputs to console (though it seems to just spit out my actual file, and not with any of the suggested fixes). I have hlint errors showing up in spacemacs:

(See the blue underlines).

However, when I run, i.e., hlint-refactor-at-point (with the cursor at one of the data's above), then the error isn't fixed (i.e., the data isn't converted to a newtype). Nothing happens. When I enable toggle-debug-on-error I don't get an error/backtrace. Nothing shows up in *Messages*. My exec-path is also configured correctly.

One thing that may be of issue is that I don't have a program called refact, but I do have a program called refactor (I think this is the binary associated with apply-refact).

NOTE: In case it matters I wasn't able to get hlint errors to show up in spacemacs without inserting this into my .spacemacs:

;; https://stackoverflow.com/questions/40400547/using-hlint-with-intero-on-emacs
(with-eval-after-load 'intero
  (flycheck-add-next-checker 'intero '(warning . haskell-hlint))
)

EDIT: After restarting emacs, hlint refactoring seems to work on everything but do blocks, where hlint seems to get tripped up about where the actual error is located (telling me its on a do block but in fact its somewhere else on the line).

[mpickering]

If I remember right the "data to newtype" hint is not one which is fixed by the plugin. Perhaps try another hint?

[listx]

I solved this by installing the refactor binary with

stack install apply-refact