Dmitriy Kalinin

@smg8fe for your use case you can use something like this for now, `echo '#cloud-conifg'; ytt -f .`

@PyMeH our thinking was to help catch mistakes by authors if they forget to add `@` after '#' (since IDEs for example dont help to add it). there is currently...

hey @groundnuty, as you found this is one of the known limitations (https://carvel.dev/ytt/docs/latest/known-limitations/); however, we should have probably made more clear that results are "undefined" (instead of providing _some_ example)....

short repro scenario: https://get-ytt.io/#gist:https://gist.github.com/cppforlife/89108d13240490f604fc98e7ac04c236

ytt has several "special" functions that need to integrate deeper into structure building stack (vs regular functions that just return structures). template.replace is one of them. will dig in further...

hey @dben, ytt has a goal to produce configuration deterministically (one example is no access to random data). bcrypt ultimately needs a salt (in addition to password and cost factor)...

closing this for now since bcrypt requires access to randomness unless salt is provided explicitly. at that point you might as well provide bcrypted password. feel free to reopen if...

this ~~related~~ is task in https://github.com/vmware-tanzu/carvel-ytt/issues/112 > traverse value for merge/replace/append/insert (since they may include overlay directives) when short circuiting

i would be interested in seeing overlay.subset() matcher extended to be able to use other matchers within it: e.g. `overlay.subset({"key": overlay.any(), "foo": lambda ...})`

> return is only allowed within a function that's still one of the options. we can rewrite ast to support return at top level. question is it a good idea?