Open Morgawr opened 6 months ago
Hmm
The best solution might be to allow dictionaries to specify the correct furigana distribution. I think that's probably the only way we'll be able to ensure the info is reliable. There's no way of figuring out the correct distribution for a word like 好き嫌い unless you know that きら.い is a standard reading of 嫌. Unless we want to load tens of thousands of kanji readings into Yomitan and make it attempt to solve these distributions every time.
There's a more thorough explanation in this jitendex bug report.
How to replicate the problem:
If you notice, yomitan applies full furigana over the whole expression because it doesn't know how to split the kana distribution between the repeated entries (double い in おねがいいたします and double き in すききらい).