Hi, I want to ask about the problem I experience with Chinese input methods. For those who doesn’t know how IM works, basically it is just inputting normal English characters and then you select the Chinese characters encoded by the alphabetical input. For example I need to input “wo” to input 我, the wo should be just an intermediate state that shouldn’t be outputted to the editor.
So I from time to time experience leaking of English inputs at the start of the line. I think I reported this problem before but I got no luck narrowing down the scope. The only information I’m certain is 1) it is a silverbullet specific problem, I cant reproduce in the codemirror 6 editor example directly 2) it only happens in an empty line and only triggers if I start by entering Chinese 3) it is cross browser problem (both in firefox and chromium) but not on mobile firefox. These syndromes makes me suspect that it is some special handling at the start of the line that breaks the input components and results in a leak.
So I want to try another way by asking where should I find those additional special handling at the start of the line? Can someone give me a brief overview on where could I find the relevant code? Even though I have no web development background I would still want to try finding the problem because it seems that no other IM users are involved right now XD. Thanks