bug#17742: Acknowledgement (Support for enchant?)

[Eli Zaretskii]

> From: Reuben Thomas <rrt@sc3d.org>
> Date: Mon, 19 Dec 2016 22:04:55 +0000
> Cc: 17742@debbugs.gnu.org
> 
> 1. Assuming my patches to enchant are accepted sooner rather than later, and 
> there's a minor Enchant release soon, then accept an initial implementation 
> of Enchant support in Emacs with a fixed casechars value. This is no worse 
> than for Aspell.
> 
> 1a. If someone wants to add a way for Emacs to parse hunspell dictionaries 
> when used via Enchant, that's fine by me as a temporary workaround.
> 
> 2. When hunspell 2 is released, hopefully there will be an official channel 
> for Emacs to get this information. Any workaround introduced as per 1a above 
> would now be more solid.
> 
> 3. Once one of Enchant's supported engines has an official way to get this 
> information, then it's a good time to add an API to Enchant too (and support 
> in the standalone binary).
> 
> Overall, there's no hurry. We have precise casechars for hunspell 
> dictionaries today (though as I mentioned elsewhere, there may still be 
> problems with using them). Enchant support for now is useful for the spelling 
> checkers it supports that Emacs does not; obviously, Emacs's direct hunspell 
> support is better for now than via Enchant. It would be nice to fix that 
> eventually and use only Enchant, but there's no need to rush.

For someone who would like to use Enchant with Hunspell as its engine,
this sounds sub-optimal to me, if 1a is not implemented as part of the
initial Enchant support, or at all.  How will we be able in good faith
to advertise support for a new speller when in fact it degrades, even
if slightly, the direct interface with that very engine?