| So I'd really like to get some feedback on two possible syntax approaches - I'm fine with either one, so I need input. =) First, which is what's currently in the PR:
agent {
label "foo"
}
agent {
docker "httpd:2.4.12"
dockerArgs "-v /tmp:/tmp -p 80:80"
label "foo"
// the below would be for things like private registry, kubernetes pod config, whatever
someComplicatedOption {
optA "banana"
optB "monkey"
}
}
agent {
dockerfile "true"
dockerArgs "-v /tmp:/tmp -p 80:80"
label "foo"
}
Second:
agent {
// i.e., DeclarativeAgent types with only one required argument and only one specified argument don't need the nested block
label "foo"
}
agent {
// Also would work with 'docker "httpd:2.4.12"' if the image is the only desired argument
docker {
image "httpd:2.4.12"
args "-v /tmp:/tmp -p 80:80"
label "foo"
// the below would be for things like private registry, kubernetes pod config, whatever
someComplicatedOption {
optA "banana"
optB "monkey"
}
}
agent {
// Also would work with 'dockerfile true' as above
dockerfile {
args "-v /tmp:/tmp -p 80:80"
label "foo"
}
}
And of course, agent any and agent none will still be working. Thoughts, Patrick Wolf, Michael Neale, James Dumay, rsandell, aurelien leboulanger, Dominik Bartholdi, Logan Glickfield, Peter Leibiger? |
