Details
-
Type:
Bug
-
Status: Closed
-
Priority:
Major
-
Resolution: Fixed
-
Affects Version/s: 10.0.1
-
Fix Version/s: 10.0.2
-
Component/s: None
-
Labels:None
Description
Function my_b_fill() in mysys/mf_iocache2.c calls my_read() directly on line 222, not through mysql_file_read() as it is done in all other places. Please check whether this is a bug and needs to be changed.
I'm not saying confidently that this is a bug because changing my_read to mysql_file_read breaks perfschema.relaylog test in a suspicious way.
Gliffy Diagrams
Attachments
Activity
- All
- Comments
- Work Log
- History
- Activity
- Transitions
I think new result (of perfschema.relaylog) is correct. Like, one cannot expect that master-index file was never read in the test, not even after show binlog events. The old result shows that it wasn't, the new results shows that is was.