Go 1.9 removed "./..." matching the vendor directory while Go 1.10 added
build and test caching. This means we no longer need to grep out
vendored matches (except for golint which doesn't implement the same
behavior), and we no longer need to pre-build packages with "go build -i".
https://golang.org/doc/go1.9#vendor-dotdotdothttps://golang.org/doc/go1.10#build
- adapted TestUnmarshalConfig to ensure the fields are read in
- added a test to see that at least MaxOpenConns works:
- this is only exposed through (*db).Stats() in go 1.11, so this test
has a build tag
- the other two configurables can't be read back, so we've got to
trust that the mechanism works given the one instance that's tested..
Signed-off-by: Stephan Renatus <srenatus@chef.io>
prior to this change, many of the functions in the ExecTx callback would
wrap the error before returning it. this made it impossible to check
for the error code.
instead, the error wrapping has been moved to be external to the
`ExecTx` callback, so that the error code can be checked and
serialization failures can be retried.
Add 'both' option to use team name AND slug in TeamNameField
This allows the connector to be configured with both so that both team names and slugs can be returned in the groups. This makes configuring teams in an application a bit more foolproof; we would often have confusion over whether the team name or slug should be given by the user, so it's easier to just allow both, since collisions shouldn't be possible anyway.
this will result in both the team name *and* the team slug being
returned for each team, allowing a bit more flexibility in auth
validation.
Signed-off-by: Topher Bullock <tbullock@pivotal.io>
Signed-off-by: Alex Suraci <suraci.alex@gmail.com>
Use pq connection parameters instead of URLs for postgres connections
This enables the use of socket paths like /var/run/postgresql for the 'host' instead of requiring TCP. Also, we know allow using a non-default port.
otherwise it's impossible to use a Unix socket, as the path gets escaped
awkwardly.
Signed-off-by: Ciro S. Costa <cscosta@pivotal.io>
Signed-off-by: Alex Suraci <suraci.alex@gmail.com>
this failed on my machine due to the unexported 'loc' field of the time
structure - it was nil in one and set to a ton of tiemzone data in the
other. instead let's just compare the unix timestamp value and zero it
out for the struct comparison.
