…t errors (#1127)

* Add `golangci-lint`, fix `gosimple`, `govet` and `unused` linter complaints

* Go 1.16

* Update copyright dates

…e commit

* compound pk tests

* more details in failure diff

* more elaborate test; the pk-ts one consistently fails

* Fix merge conflict

Co-authored-by: Shlomi Noach <[email protected]>
Co-authored-by: Tim Vaillancourt <[email protected]>
Co-authored-by: Tim Vaillancourt <[email protected]>

Co-authored-by: dm-2 <[email protected]>

…)` (#1135)

* Use `switch` statements for readability

* Simplify initBinlogReader()

* Add context/timeout to HTTP throttle check

* Dont run `.GetThrottleHTTPInterval()` on every loop

* Update help message

* Var rename

* 2022

* Add timeout flag

* Add unix/tcp server commands, use ParseInt() for string->int64

* Var rename

* Re-check http timeout on every loop iteration

* Remove stale comment

* Make throttle interval idempotent

* var rename

* Usage grammar

* Make http timeout idempotent too

* Parse time.Duration once

* Move timeout to NewThrottler

* Help update

* Set User-Agent header

* Re-add newline

Co-authored-by: dm-2 <[email protected]>

* Cancel any row count queries before attempting to cut over

Closes #830. Switches from using `QueryRow` to `QueryRowContext`, and
stores a context.CancelFunc in the migration context, which is called to
halt any running row count query before beginning the cut over.

* Make it threadsafe

* Kill the count query on the database side as well

* Explicitly grab a connection to run the count, store its connection id
* When the query context is canceled, run a `KILL QUERY ?` on that connection id

* Rewrite these to use the threadsafe functions, stop exporting the cancel func

* Update logger

* Update logger

Co-authored-by: Tim Vaillancourt <[email protected]>
Co-authored-by: Tim Vaillancourt <[email protected]>
Co-authored-by: dm-2 <[email protected]>

…istics (#844)

* Check RowsAffected when applying DML events to get more accurate statistics

Addresses #600.

When applying a DML event, check the RowsAffected on the `Result`
struct. Since all DML event queries are point queries, this will only
ever be 0 or 1. The applier then takes this value and multiplies by
the `rowsDelta` of the event, resulting in a properly-signed, accurate
row delta to use in the statistics.

If an error occurs here, log it, but do not surface this as an
actual error .. simply assume the DML affected a row and move on. It
will be inaccurate, but this is already the case.

* Fix import

* update wording to warning log message

Co-authored-by: Tim Vaillancourt <[email protected]>

The base image is `buster` instead of previously assumed `stretch`