nolar / kopf / 1134

82%

main: 91%

push

travis-ci

WIP: functional auto-discovery of resources & namespaces w/ glob matching

For namespaces: `-n default -n kube-*` or `-A`
For resources: `@kopf.on.create('*', '*', '*')`

Play with:

* `kopf run ... -n default -n kube-*`
* `kubectl create -f examples/crd.yaml`
* `kubectl create ns kube-xyz`
* `kubectl delete ns kube-xyz`
* `kubectl delete -f examples/crd.yaml`

TODOs:

- support built-in resources for globs (via scan_resources() listing)
- support short forms of handler decorators (short/singular names only)
- warning for no -A or -n (or current context should be used?)
- remove all resource info caches from the aiohttp session/context
- better docstrings in discovery.py
- registries: either backward compat, or drop the legacies already (too hard).
- multi-namespace peering (consider partially overlapping operators?)
- tests
- docs & examples

909 of 1241 branches covered (73.25%)

Branch coverage included in aggregate %.

3352 of 3987 relevant lines covered (84.07%)

7.52 hits per line