Speed-up dbcs.MatchComment.

[chhsiao1981]

Provide the github link(s) of the file#line (請提供 github 關於 code 的 link)

https://github.com/Ptt-official-app/go-openbbsmiddleware/blob/main/dbcs/match_comment.go#L56

What would you like to discuss (你想要討論什麼呢？～)

Currently we loop-through all the comment-types and then determine the comment-type with the smallest index.
However, it's highly possible that some comment-types never exist in the whole content, ending-up O(n), where n is number of the bytes in the content.

It's better that we loop-through the content and check whether the byte is the starting point of some comment-type.
Given that we do MatchComment a lot in commentDBCS, this should improve speed significantly.