Arch Linux

The hunspell-en package includes many variants of the english dictionaries, however many of those are provided by either en_US or en_GB via symlinks, with the actual dictionary, hyphenation and thesaurus files being only en_US, en_GB and en_CA. This causes the appearance of a big list of dictionaries in programs that use hunspell.

The feature request is to split the hunspell-en package into hunspell-en-base and hunspell-en (or other names deemed appropriate).

hunspell-en-base would provide en_{CA,GB,US}.{aff,dic} in /usr/share/hunspell and th_en_US_v2.{dat,idx} in /usr/share/mythes and the respective symlinks in /usr/share/myspell/dicts.

hunspell-en would depend on hunspell-en-base and provide the symlinks adding all the variants.

By keeping the hunspell-en name for the package that provides all the variants, there should be no visible change, breakage or need for intervention for current users of the variants and it would still be intuitive to find. The users that don't need the variants can probably figure things out for themselves and install only hunspell-en-base.

I know this issue was brought up before [2] and closed as wontfix, as the whole list of variants is intended, however this would allow the users who don't need the variants to have a reduced number of listed dictionaries, avoid suggestions of manually deleting things from the filesystem as in [1] and keep everything up to date thus avoiding possible breakage in the future.

[1] https://bbs.archlinux.org/viewtopic.php?id=144299
[2] https://bugs.archlinux.org/task/24947