Hrm, this is tricky. They have separate entries in the EDICT dictionary (which the site's dictionary is based on), so without a significant amount of extra checks, linking the two of them would result in a new copy of the "secondary" one every time I do an update from the EDICT. This one will unfortunately have to remain until if/when I change up that part of the import system.
I think something went wrong... now the "alternate kanji" entries for each term are just duplicates of themselves instead of having an entry each for 齣_り / コマ_り / こま_り like they did before.