Right now we use pg_regress to test the extension. But given the fact that they're simply comparing the output of a test run with their expected output, they are version-dependent (some PostgreSQL versions are more chatty about certain things than others), it's kinda hard to see what actually failed on first glance (the diff first shows all the stuff that wasn't output before printing any error messages).
pg_tap is a unit testing extension for PostgreSQL. So instead of comparing test output, it'll actually look at the data (and has helper functions that describe what we're testing for).
see #2
When that test migration is done, it might be interesting to think about something like travis-ci.
Right now we use
pg_regressto test the extension. But given the fact that they're simply comparing the output of a test run with their expected output, they are version-dependent (some PostgreSQL versions are more chatty about certain things than others), it's kinda hard to see what actually failed on first glance (the diff first shows all the stuff that wasn't output before printing any error messages).pg_tap is a unit testing extension for PostgreSQL. So instead of comparing test output, it'll actually look at the data (and has helper functions that describe what we're testing for).
see #2
When that test migration is done, it might be interesting to think about something like travis-ci.