Closed df8s7d9g8s closed 9 months ago
I considered that at some point, but never looked into it. My main concern is hash collisions (and potentially somewhat higher resource usage). I'm not against it but overall I don't think it's worth doing the change for a few duplicates.
Don't know how it is for larger instances like review.px4, but at least for the one I host for my organization I keep having people upload the same log repeatedly instead of making public reports and using them, so I thought it would be a general concern more than a "few duplicates".
Is there a reason why uuids are generated with
uuid.uuid4()
instead of something likeuuid.uuid3(uuid.NAMESPACE_DNS, str(open(filename, "rb")))
?Would prevent getting multiple URLs and database entries when uploading the same log file and redirecting us to the existing log which is how I expected it to behave.