diego dupin

error in travis is because of a different error (common.fakeServerPort 32893 was already used) (FakeServer test would probably be better using port 0, then getting resolved port)

You can either set a global query timeout for your connection using `set @@max_statement_time=XX` (be carefull, value for MariaDB are in seconds, and MySQL in milliseconds) if you are using...

It is number of operation per second (higher is better). If needed here is the code used for benchmark : https://github.com/rusher/connector-benchmark/blob/main/scripts/dotnet/Program.cs Each connector has different problem, like some needs to...

> If your benchmarks are primarily driven by latency, then I would expect sync to look better than async, and async only start to shine in a highly concurrent benchmark...

rebase on current master, and corrected the point above

rebase on current master, and added https://github.com/go-sql-driver/mysql/pull/1696/commits/fde44b6541b7e93a427aca029662657c383f7969 to ensure public key casting

Do you have any thoughts or concerns about when this can be merged?

maybe a first step would be to make that not pluggable ( i mean not public ) ?

The problem with current correction is that the expected number of columns is returned, but no columns follows. see https://mariadb.com/kb/en/com_stmt_prepare/ : for you example, after "0c 00 00 01 00...

force push implementation with performance improvement. current implementation : ``` Running tool: /usr/bin/go test -benchmem -run=^$ -bench ^BenchmarkInterpolation$ github.com/go-sql-driver/mysql goos: linux goarch: amd64 pkg: github.com/go-sql-driver/mysql cpu: 11th Gen Intel(R) Core(TM)...