This shows issue 8 isn't caused by a simple bug in clj-record. Hopefully the reporter can provide more details if he's still seeing this issue.
…f obsoleted by switching to lein.
…r of resetting the whole database before running lein test.
…to and has-many.
…cords previously had.
…lbacks which made sense for clj-record.
…s with underscores when used as table names and foreign key column names.
Having the sole source path on the classpath when running tests had caused me not to realize that certain namespaces weren't being compiled. The test target of the build now puts the compiled classes directory on the classpath, as well as the new /test source directory, but not the /src directory.