I think that in the ls->read workflow, Nextcloud shouldn't normalize the response from SMB and should issue back to SMB whatever SMB returned to Nextcloud.
According to Unicode, it should be allowed to and the SMB server should be able to handle it. That's kind of the point of normalization, they're meant to be done before all comparisons so that exactly this doesn't happen. Your suggestion is just premature optimization, i.e. eliminating a redundancy.
Unicode doesn't say anything about what "should be allowed to" with respect to an unrelated protocol. If the protocol says that filenames are sequences of 16-bit values that have to be compared one by one, then that's what it is.
If comparisons are being made of Unicode strings, sure. Does the protocol actually defines the identifier in question as a Unicode string, though? Or as an array of 16-bit ints?
