Yes and no.
Yes, we should not use "_" and not translate it in the code. We call it a bug.
But i dont think the result will be better. Cause, i am sorry, but i dont know how we write "english" in chinese. Then it will be a mess if you click somewhere you should not. But i hope chinese people know how to right "chinese" in english (in msgid if you prefered). Then i dont know if we should apply this patch; but at least we should add a comment about the result of this "_" skip.
The real fix for languages should be to use languages name in this language itself. I mean chinese in chinese. i.e. wikipedia. I start a long time ago working on a script to create a small font file containing all characters need to display all language (chinese in chinese...) but i dont find a script based tool to extract and create font. If you know something like that, i am interesting to work on that. (idea is to use a few memory, and not generating all unicode page in memory, cause i think it will not be nice).