This PR allows the LocalCluster library to work with any node name instead of the hardcoded manager@127.0.0.1. This may be useful if other parts of the test suite require a specific basename or externally reachable hostname. In my case, the requirement was for the test runner node to be reachable from Docker nodes that are needed for integration testing (these nodes run a completely different pre-built Elixir project so LocalCluster is still needed for testing clustering of the current project).
This is not a breaking change in any way since manager@127.0.0.1 is still the default and the dynamic host lookup (that allows this lib to work properly even if :net_kernel.start is invoked before & outside of LocalCluster.start) will yield the same value as the previously hardcoded one.
I've provided a test case that requires a custom host in /etc/hosts but the test case nicely explains it all and it reverts to the original default setup.
This PR allows the LocalCluster library to work with any node name instead of the hardcoded
manager@127.0.0.1
. This may be useful if other parts of the test suite require a specific basename or externally reachable hostname. In my case, the requirement was for the test runner node to be reachable from Docker nodes that are needed for integration testing (these nodes run a completely different pre-built Elixir project so LocalCluster is still needed for testing clustering of the current project).This is not a breaking change in any way since
manager@127.0.0.1
is still the default and the dynamic host lookup (that allows this lib to work properly even if:net_kernel.start
is invoked before & outside ofLocalCluster.start
) will yield the same value as the previously hardcoded one.I've provided a test case that requires a custom host in
/etc/hosts
but the test case nicely explains it all and it reverts to the original default setup.