Use env vars to be explicit about the model. (Fixes #178)
The placement of the CLI arg for specifying the model is complicated by the fact that Juju 1 has compound commands (such as juju action do) while other commands have to have the arg as early as possible (e.g., juju scp passes later args through to scp). The env vars will override the active environment from the state file.
This also adds functional testing to the Travis config for Juju 2. I'm not sure of a clean way to force it to use juju-1 even if we install that, but I'm open to ideas.
Use env vars to be explicit about the model. (Fixes #178)
The placement of the CLI arg for specifying the model is complicated by the fact that Juju 1 has compound commands (such as
juju action do) while other commands have to have the arg as early as possible (e.g.,juju scppasses later args through to scp). The env vars will override the active environment from the state file.This also adds functional testing to the Travis config for Juju 2. I'm not sure of a clean way to force it to use
juju-1even if we install that, but I'm open to ideas.