PURGE BINARY LOGS not raising error when not being able to purge a file still being read by a slave

The KB page on PURGE BINARY LOGS says:

If a replica is active but has yet to read from a binary log file you attempt to delete, the statement will fail with an error.

But actually purging just stops at that file without an error or warning.

Not sure whether we should consider this an actual bug or "just" a documentation error, but I'd like to actually have PURGE BINARY LOGS generate a warning in this case, not an error.

With that we would not break backwards compatibility by suddenly make this fail, but at the same time we would give feedback to the user about less files having been purged than requested.