Status: Closed (View Workflow)
Affects Version/s: 5.5, 10.0, 10.1, 10.2.13, 10.2, 10.3
Component/s: Character Sets
Environment:Darwin crest.fritz.box 17.4.0 Darwin Kernel Version 17.4.0: Sun Dec 17 09:19:54 PST 2017; root:xnu-4570.41.2~1/RELEASE_X86_64 x86_64
Changing the default character set to utf8mb4 changes query evaluation in a very surprising way. (Please feel free to set a better title).
This is actually a followup bug report to this: https://bitbucket.org/zzzeek/sqlalchemy/issues/4222/query-yields-different-result-via-sqla (see there for more detail)
Here's the problem: This query: ```select seq, replace(uuid(), "-", "") from seq_0_to_9;``` has a completely different result, wether you connect to mysql with a character set of utf8 or utf8mb4.
Here's an example:
This returns 10 UUIDs that are all the same. *I believe this to be wrong*
Calling that same query with utf8 as the format yields 10 different uuids - which I believe to be correct:
This also happens on on at least on current Fedora. (I can look up the details tomorrow)
On a further note, this result is