* wip
* made progress on new SQLite type stuff
* we're getting places
* largely working now
* tidying up
* almost there
* time to deal with failing tests
* seemed to have fixed sqlserver issue
* seems to all be fixed
* minor tidy up
* debugging
* generic retry functionality; some additions to stringz.go
* fiddling with magefile_sakila
* added test lock function testh.Lock
* bunch of linting, goimports, what not
* upgraded mattn/go-isatty version because it was importing a incompatible version of golang.org/x/sys/unix
* changed parsing of SQL Server DSN due to changes to url.ParseQuery no longer accepting semicolon
* updated xo/dburl pkg to v0.11.0
* removed dead underscore assignment
* updated GH workflow go action to 1.19
* implementation work for json importers
* json driver checkpoint
* working on json.ParseObjectsInArray
* json.ParseObjectsInArray seems to be working
* checkpoint while tidying up ParseObjectsInArray
* more tidy checkpoint
* more tidy checkpoint 2
* tidying up ParseObjectsInArray
* tidy up
* code/docs cleanup
* more cleanup of json driver
* more cleanup of json driver
* flat json import seemingly working
* improvements to json driver
* json writer now prints empty [] for postgres empty tables
* Type Detector refactor
* json driver impl; refactoring of source.Files reader et al
* working on kind detector
* significant switcheroo of packages
* partway throught refactoring Kind
* major package switcheroo for Kind
* improvements to sqlite source metadata query
* table metadatawriter colorization tuning; minor testh.Helper refactoring
* cleanup of metadata for multiple drivers
* more cleanup of source/table metadata
* yet more cleanup of source/table metadata
* improvements to mysql SourceMetadata
* improvements to mysql TableMetadata
* yet more fiddling with mysql metadata
* added Go dependency cache to go.yml workflow
* more windows path issues
* docs
* restricting workflow triggers to master
* wrong cleanup order for downloaded files
* de-restricting workflow triggers, for now
* more out-of-order cleanup issues that only manifested on windows
* more out-of-order cleanup issues
* removed mediatype check from TestFetchFile; FetchFile is deprecated and not worth fixing
* tracked down out-of-sequence close/remove issue with sqlite
* probably long-running test failure
* disabling long-running tests that are breaking CI
* switching to go test -short because tests seem to be taking too long in CI
