Last modified: 2014-10-07 05:43:05 UTC
I think will be good idea to generate upload log entry when duplicate file uploaded. I this case such uploads will be visible to admins and automatic bot processing will be possible without traversing all files. See also bug 14171 for other useful application of such logging.
How do you define a "duplicate"? By checksum? Which exact "upload log" (URL) does this refer to?
I think checksum is best way to find duplicate. I meant type=upload for log.
Could duplicate file detection be done in AbuseFilter, which would allow logging, and pre-programmed management? It appears it has the sha as a variable. See https://www.mediawiki.org/wiki/Extension_talk:AbuseFilter/Rules_format#file_sha1