When trying to publish a translation with content that triggers an existing abuse filter, the captcha returns an error even if answered correctly. This prevents the user from publishing giving a reason (wrong captcha provided that is misleading. The user should be able to pass the captcha and get the abuse filter error with the details on what is wrong.
I tried this on the testing servers by adding text "hahahahahahahahahahahahaha" after a paragraph in its own line, which should trigger the "repeated characters" abuse filter defined in T192524.
Related tickets: