Unexpected capital letters returned for certain capitalized misspellings about nspell HOT 3 CLOSED

Awesome, thanks for digging in and doing work on these two PRs!

I see that both PRs are WIP: what can we do to make them actionable? How can I help?

from nspell.

I see that both PRs are WIP: what can we do to make them actionable? How can I help?

In relation to a new issue I've raised, check out updates I've made on PR #39.

I think it's almost ready!

from nspell.

I see that both PRs are WIP: what can we do to make them actionable? How can I help?

The most helpful feedback you could give right now would be how I should approach the test cases:

• Is it OK if the tests that fail are essentially the same except that the suggestions are in a different order?

After that, I'd like to hear your thoughts on what seems to be broken keyboard groups logic. For example, the input Ghandi leads to an initial "edit" of Grandi even though h and r are not in the same keyboard group. Is it even important that the keyboard groups do not work as expected? Essentially, if that issue were fixed (like in PR #39), we would not need a separate fix for this capitalization issue.

I've only set the WIP status because the tests are failing. If backwards compatibility for certain arbitrary test cases is not an issue, then we would just need to update the tests. PR #38 is failing the tests less severely than PR #39, though I think PR #39 makes more logical sense.

I've spent some time today looking into how to make PR #39 match the tests more closely, and it really seems that one test case in particular (test #46) is really reliant on replacements made by what seems like a buggy implementation of replacements based on keyboard groups.

from nspell.