[Haskell-cafe] emacs problem : haskell mode keeps hanging to report types, renders emacs useless

Sean Matthews seanmatthews1 at gmail.com
Thu Nov 23 22:11:28 UTC 2017


Alas I spoke too soon, and my problem is still there.  So it is back to the
cafe, looking for suggestions.  Here is my problem in more detail:

 - I am running a windows system. (note Windows, not Unix, Linux, OS X, or
any other Unix-a-like).

 - I have a clean emacs installation - only thing I have done is to install
haskell-mode using the install-package function.

 - I start-up interactive haskell mode (C-c interactive-haskell-mode, C-c
C-l (and I get a parsing complaint, but we ignore that)).

[at this point I have a live haskell process that I can interact with]

 - I load my file (not large) into haskell.

 - Emacs starts persisting in telling me in the Echo region the type of the
term under the cursor - all the time - and locks down to do this.

 - The type report uses unicode text to do this (note that if I ask for the
type with C-c C-t or C-c C-i, I also get the type, but using ascii
characters, so it does not appear to be the same functionality that is
being triggered - if I turn off haskell-doc then the problem continues).

Anybody any suggestions, recognise this problem, or even just an address
where I can submit a bug report?
I cannot be the only person trying to work with haskell mode on windows.

Thanks

Sean Matthews

On Sun, Nov 5, 2017 at 8:24 PM, Sean Matthews <seanmatthews1 at gmail.com>
wrote:

> Problem fixed. Not an interesting problem, either. Turns out that I had
> not quite completely cleaned out my emacs before reinstallation.
>
> Many thanks,
>
> Sean
>
> On Sun, Nov 5, 2017 at 10:46 AM, Saurabh Nanda <saurabhnanda at gmail.com>
> wrote:
>
>>
>>> 1. Exact behavior is that Emacs freezes for about a minute and displays
>>> (in bold) in the type of the word I'm over.
>>>     Thus, I was trying to edit a small a-list with dummy 0.000s in the
>>> range.  Every third keystroke, emacs stopped to tell me that
>>>     my cursor was over 0.000::Double and I have to wait an indeterminate
>>> time (at least tens of seconds, sometimes a minute or more)
>>>     to get control back.
>>>
>>
>>
>> So, the underlying mechanism to determine type at current cursor position
>> isn't very fast. In spacemacs+intero C-c C-t (or C-c C-i) takes anywhere
>> between 1 to 10 sec for me. Something in your setup has made this lookup
>> automatic. It's trying to determine the type as your browse around your
>> source code. Do you have any idea what could have caused that?
>>
>> You could try one more thing. Try delivering a SIGUSR2 to the Emacs
>> process when it's stuck. It will stop doing whatever is holding it up and
>> drop you into some sort of a debugger. That can give you some clues about
>> what is causing this behaviour in your setup.
>>
>> -- Saurabh.
>>
>
>
>
> --
> Sean Matthews
> seanmatthews1 at gmail.com / +49 1515 800 1901 <+49%201515%208001901>
>



-- 
Sean Matthews
seanmatthews1 at gmail.com / +49 1515 800 1901
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20171123/259f6798/attachment.html>


More information about the Haskell-Cafe mailing list