Use OpenOffice sources for Armenian dictionaries

[hubgit]

As the old Arak-29 site no longer exists, and the new site doesn't contain any spell checkers, fetch the Armenian dictionaries from the OpenOffice extensions instead, which contain the same data.

[wooorm]

Why would it be better to get unmaintained libraries through one more source?

Did you see https://github.com/wooorm/dictionaries/blob/f3bba386c50ba91805eccda869af5d915808d9ce/script/crawl.sh#L164C9-L167 btw?

[hubgit]

I guess it depends whether it should always be possible to build all the dictionaries from source, or if you're happy for this repo to contain some dictionaries that aren't available anywhere else (which would be fine).

One advantage of using the new source would be that it's easier to check for and pick up updates, if it's updated from another source in the future.

I don't mind either way, happy to close this if you'd prefer to.

[hubgit]

Did you see https://github.com/wooorm/dictionaries/blob/f3bba386c50ba91805eccda869af5d915808d9ce/script/crawl.sh#L164C9-L167 btw?

I did see the comment, but I suspect I may not have fully understood it 🙂

I took it to be an explanation of why the current dictionaries were chosen, but it sounds like it might actually all be part of the TODO for future improvements?