mbox

[bug#36087,0/9] Add further Aspell dictionaries.

Message ID 87tvd96177.fsf@zete.tk
Headers show

Message

Jens Mølgaard June 1, 2019, 10:37 a.m. UTC 
Hi Guix,

I have added some additional Aspell dictionaries. I wasn't sure whether
to separated them out or not, as they're not quite stand-alone packages,
so I'll send them as a patch series - squish them together in one commit
if that makes more sense.

In case I or someone else wants to add them in the future, some older
dictionaries such as Icelandic and Norwegian Bokmål contain non-utf8
filenames and will fail to build. Ludovic suggested adding glibc-locales
as input and a pre-unpack phase to do (setlocale LC_ALL
"en_US.iso88591").

--
Jens Mølgaard

Jens M (9):
  gnu: Add Arabic Aspell dictionary.
  gnu: Add Belarusian Aspell dictionary.
  gnu: Add Danish Aspell dictionary.
  gnu: Add Finnish Aspell dictionary.
  gnu: Add Hindi Aspell dictionary.
  gnu: Add Maori Aspell dictionary.
  gnu: Add Norwegian Nynorsk Aspell Dictionary.
  gnu: Add Portuguese Aspell Dictionary.
  gnu: Add Ukrainian Aspell Dictionary.

 gnu/packages/aspell.scm | 69 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 69 insertions(+)

Comments

Ludovic Courtès June 5, 2019, 3:11 p.m. UTC | #1 
Hello,

Jens Mølgaard <jens@zete.tk> skribis:

> I have added some additional Aspell dictionaries. I wasn't sure whether
> to separated them out or not, as they're not quite stand-alone packages,
> so I'll send them as a patch series - squish them together in one commit
> if that makes more sense.

I’ve applied them all and adjusted the commit log to follow our
conventions.  Thanks!

There’s a minor issue: the URL computed for the pt-pt dictionary is
incorrect, because the URL should read “pt_PT” and not “pt-pt”.  It
turns out that the tarball was found on the tarballs.nixos.org
content-addressed mirror, so it builds flawlessly.  However, it’d be
nice to adjust ‘aspell-dictionary’ so that it can compute the right
upstream URL for edge cases like this one.

Would you like to give it a try?

> In case I or someone else wants to add them in the future, some older
> dictionaries such as Icelandic and Norwegian Bokmål contain non-utf8
> filenames and will fail to build. Ludovic suggested adding glibc-locales
> as input and a pre-unpack phase to do (setlocale LC_ALL
> "en_US.iso88591").

Future work!  :-)

Thanks,
Ludo’.
Jens Mølgaard June 6, 2019, 12:25 a.m. UTC | #2 
Sure! I'll have a look...

Ludovic Courtès <ludo@gnu.org> writes:

> Hello,
>
> Jens Mølgaard <jens@zete.tk> skribis:
>
>> I have added some additional Aspell dictionaries. I wasn't sure whether
>> to separated them out or not, as they're not quite stand-alone packages,
>> so I'll send them as a patch series - squish them together in one commit
>> if that makes more sense.
>
> I’ve applied them all and adjusted the commit log to follow our
> conventions.  Thanks!
>
> There’s a minor issue: the URL computed for the pt-pt dictionary is
> incorrect, because the URL should read “pt_PT” and not “pt-pt”.  It
> turns out that the tarball was found on the tarballs.nixos.org
> content-addressed mirror, so it builds flawlessly.  However, it’d be
> nice to adjust ‘aspell-dictionary’ so that it can compute the right
> upstream URL for edge cases like this one.
>
> Would you like to give it a try?
>
>> In case I or someone else wants to add them in the future, some older
>> dictionaries such as Icelandic and Norwegian Bokmål contain non-utf8
>> filenames and will fail to build. Ludovic suggested adding glibc-locales
>> as input and a pre-unpack phase to do (setlocale LC_ALL
>> "en_US.iso88591").
>
> Future work!  :-)
>
> Thanks,
> Ludo’.