Dr. Stefan Schimanski
Dr. Stefan Schimanski
Ack. Leaving final review to @soltysh
Note that you break users following docs on the web showing "kubectl crossplane install". Even if it is just a shallow wrapper, breaking users is not good. Deprecation of the...
@Harvie I think travis build with 2.7 is broken independently from this PR.
I guess the generators were never tested on Windows. PRs fixing this are very welcome.
I ported the Alloy pre-compile plugin to stock Titanium SDK: https://github.com/sttts/ti-commonjs-wrap-plugin
This needs a proof PR against Crossplane. I have the suspicion that it breaks the composite reconciler as it relies on the inconsistent, additive behaviour.
> applicator does not have an optimistic concurrency issue at the moment. Is that understanding wrong? It has a consistency issue though :) It does not apply what you tell...
Addressed comments. Also extended tests.
> The logic in the composite controller (also in the claim controller) heavily relies on additive behavior currently. Hence, it wouldn't be possible to switch to the patching applicator introduced...