[PATCH 3 of 4] crecord: don't abort on Windows when curses isn't available

[Matt Harbison]

On Sat, 21 Mar 2015 20:59:00 -0400, Pierre-Yves David  
<pierre-yves.david at ens-lyon.org> wrote:

>
>
> On 03/20/2015 11:11 PM, Matt Harbison wrote:
>> # HG changeset patch
>> # User Matt Harbison <matt_harbison at yahoo.com>
>> # Date 1426908913 14400
>> #      Fri Mar 20 23:35:13 2015 -0400
>> # Node ID 75ed9687553578e1253baf6b357cbe90076969a7
>> # Parent  ce3e2b1febb53e893943bae8582d1d75aca48da8
>> crecord: don't abort on Windows when curses isn't available
>>
>> This is the hack portion of the previous patch, which can be backed out  
>> once we
>> figure out how to deal with curses.  Without this, the tests affected  
>> by the
>> import problem mentioned in the previous patch fail in exactly the same  
>> way, but
>> with the nicer 'module not available' error.
>
> Wait what? you mean that mercurial have and hard dependant to curses now  
> ? If true this mean we should change things to only requires curse when  
> we actually needs it.

I guess it's a hard dependency.  IDK how much more massaging Laurent is  
doing on this code, and I'm not sure how to change that anyway.  Since the  
stack traces appeared to point to localrepo as near the beginning of the  
import chain, I'm surprised anything ran:

     'localrepo' -> 'subrepo' -> 'cmdutil' -> 'crecord'

I'm just looking to get the Windows tests running again in the short term.

Is bundling wcurses an option?  I did some quick search for it last night,  
but I've never used it before, and IDK what state it is in.

The tests actually run on Windows without it installed (with this series),  
so the tests aren't covering something (probably because it isn't  
interactive?)  But that seems like a big problem if something in core  
isn't available everywhere, and the tests don't detect it, no?

--Matt