Closed nicolasmaia closed 4 years ago
That's a good idea!
This seems fixed from my end. Did you tweak with it?
Oh, no I don't know what happened there. I mean, I half rewrote the de-inflection code so that probably fixed it but actually I didn't expect the behavior to change so I'm afraid I might have broken other things in the process.
git bisect
says this was fixed by fa1c7f77f826c0342d7aa82da3a230050f67ad2c but I have no idea why.
I'm closing this for now, but I really don't know exactly what fixed it.
This seems to be broken again.
Should be fixed in the next version.
When parsing ローマじゃないです, Rikaichamp thinks it's looking at ローマ字. This problem can maybe be fixed if you tell the program to never separate じ from ゃ (and other similar cases).