Bug#647325: Spell files include misspelling "interpretor"

[James McCoy]

On Wed, Nov 2, 2011 at 12:44 PM, Josh Triplett <josh at joshtriplett.org> wrote:
> On Wed, Nov 02, 2011 at 12:17:54PM -0400, James McCoy wrote:
>> On Tue, Nov 01, 2011 at 06:20:46PM -0700, Josh Triplett wrote:
>> > On Tue, Nov 01, 2011 at 08:13:14PM -0400, James McCoy wrote:
>> > > On Tue, Nov 01, 2011 at 02:17:21PM -0700, Josh Triplett wrote:
>> > > > The spell files used by ":set spell" do not flag the misspelling
>> > > > "interpretor".  Every dictionary I've checked either doesn't list it or
>> > > > lists it only as a misspelling of "interpreter".
>> > >
>> > > They're based on hunspell,
>> >
>> > Ah, excellent.  Do updates from hunspell occur manually, at build time,
>> > or at runtime?  (That will determine whether I just need to reassign
>> > this bug against hunspell, or whether I also need to request a rebuild
>> > of vim against the new hunspell.)
>>
>> Neither, actually.  The vim packages only ship the english spellfiles,
>> as provided in the upstream source.  Any other language is downloaded,
>> if the user allows, as described in ':help spellfile.vim'.
>
> Oh.  So, what's the procedure for fixing an issue like this in one of
> the english spellfiles?

Fix it in upstream hunspell and it will eventually be rolled into the
Vim sources by the maintainer of Vim's English spellfile.

>> We had done some work a few years ago to make language-specific
>> spellfiles packages, but at the time none of the maintainers had
>> machines that could actually build all the spellfiles.  The process
>> typically ran afoul of the OOM killer.  As such, we asked for the
>> removal of those packages (#471285).
>>
>> I might actually have a system that can build those files now, so I
>> could potentially revisit that.
>
> Seems preferable to finish the work you previously mentioned to make vim
> use enchant and read the original dictionaries directly, rather than
> working on generating vim spellfiles.

Agreed, as long as that ends up being actually feasible.  It's been a
while since I've worked on that patch set, but I seem to recall there
was a bit of bleed over of the internal spell checking to other
components (like highlighting) which could make the change difficult.

-- 
James
GPG Key: 1024D/61326D40 2003-09-02 James McCoy <jamessan at debian.org>