Error overlays

[deepfire]

No description provided.

[deepfire]

Origin PR: #766

[geraldus]

Can you please add second whitespace after . (emacs lint produces a warnings about this)?

[geraldus]

very nice!

[deepfire]

@geraldus, pushed, but, if I'm not wrong the lack of a double-space in that docstring wasn't introduced by me.

[gracjan]

@deepfire: Good stuff. There are some things to get in shape, overall this is pretty cool thing!

[deepfire]

@gracjan, I'm thinking.. is it time to separate overlay functions into haskell-overlay.el ?

[deepfire]

@gracjan, after a litany of curses I have made ghc-mod to work again -- without error-overlays, that is.
I didn't manage to get it working properly, which is a perennial issue -- and indeed, one of my chief motivators to jump ship -- all I got to work was a misplaced error overlay at the top of the file -- as in, it is misplaced, but it correlates with what GHC returns for the file.

And I can happily report that:

1. error-overlays, with haskell-process-load-file does not seem to make this user experience any worse
2. M-n/M-p doesn't seem to conflict with any ghc-mod's bindings -- and a glance at their source suggests that this is how it should be, indeed
3. the overlays from both packages coexist, which raises a certain question..

I'm not sure if this constitutes sufficient testing, though.

[gracjan]

Thanks.

[geraldus]

Cool, thanks! Let's try this out!

[deepfire]

@geraldus, it's somewhat incomplete without the other pull requests, I'm afraid..

[deepfire]

@geraldus, at a minimum you would want:

(pushnew "-Wall" haskell-process-args-ghci)

[DanielG]

Aren't these going to conflict with ghc-mod?

[deepfire]

@DanielG, do you mean with:

(defvar ghc-previous-key    "\ep")
(defvar ghc-next-key        "\en")

I must say I overlooked this, then..

But then, there's the more pressing question of having two sets of overlays on at the same time, which is still unresolved.