Bradley Grainger

Results 329 comments of Bradley Grainger

> If you are interested in profiling this to see where it's so slow I am. This library should be faster than Oracle's MySQL Connector/NET for all use cases, except...

Yes. Thanks! I think the implementation should be fairly straightforward in `SchemaProvider.cs`, with a little change in `MySqlConnection.cs` to pass the restrictions through.

> Not sure how you want to resolve this as this appears to be something specific to this data set and I am not sure I can reproduce this in...

> One specific database has the offending information in INFORMATION_SCHEMA's UPDATE_TIME but since the dates are consistent and well-formed, I don't suspect corruption, just... MySQL weirdness. Thanks; this was the...

[This run](https://dev.azure.com/mysqlnet/MySqlConnector/_build/results?buildId=2249&view=logs&jobId=f2e5acd6-ca0f-5e5f-02b7-c84ea333c209&j=0f863525-2446-520a-eaaa-52cf7f5cee55&t=5aa15591-29b4-504f-eb18-47197fa3d1cf) failed with the following exception: ``` [xUnit.net 00:00:04.32] SideBySide.SchemaProviderTests.GetSchema(schemaName: "Partitions") [FAIL] Skipped SideBySide.CommandTimeoutTests.CommandTimeoutResetsOnReadSync [1 ms] Failed SideBySide.SchemaProviderTests.GetSchema(schemaName: "Partitions") [1 s] Error Message: System.FormatException : Couldn't interpret '1899-00-�5 �5:�5:�5'...

No, it seemed like a spurious build failure unrelated to the code that was pushed. So possibly some kind of bug in MySQL Server?

[Another run](https://dev.azure.com/mysqlnet/MySqlConnector/_build/results?buildId=2298&view=logs&jobId=0f863525-2446-520a-eaaa-52cf7f5cee55&j=0f863525-2446-520a-eaaa-52cf7f5cee55&t=5aa15591-29b4-504f-eb18-47197fa3d1cf) failed (using MySQL 8.0.15 on Windows as the server): ``` [xUnit.net 00:00:04.79] SideBySide.SchemaProviderTests.GetSchema(schemaName: "Tables") [FAIL] Failed SideBySide.SchemaProviderTests.GetSchema(schemaName: "Tables") [1 s] Error Message: System.FormatException : Couldn't interpret '1899-00-�5 �5:�5:�5'...

Searching for "1899-00-É5 É5:É5:É5" finds a couple of hits in Google. `É` is `0xC9`, which could possibly be 201 or -55 as a numeric value.

I have a theory. In 8.0.15, the code to convert a date to a string is here: https://github.com/mysql/mysql-server/blob/mysql-8.0.15/mysys/my_time.cc#L1096-L1176 Note the comment: > This functions don't check that given MYSQL_TIME structure...

Just double-checking: you don't have any way to create a repro for this problem; it's just something you observe with the schema for a few (?) tables in one particular...