Chris Dostert
Chris Dostert
@apexskier thanks for filing! Just wanna clarify, is the intent only to use these creds to pull the root op being run/installed (github.com/Remitly/parakeet#v1.0.0/parakeet-ops/update-all-locales in your example) and not arbitrary descendant...
Also, per https://github.com/opctl/opctl/issues/745 just thought i'd mention, another possibility could be to allow git cred store to be a default source
Also worth mentioning, pkgs do get cached; you shouldn't be prompted more than once per pkg, but I know that doesn't solve the underlying thing you're after.
There's been the idea of a `space` floating around for some time too. Idea was to solve for at least configuration of op inputs across all the things. So your...
Right, the reason you don't get prompted when private sub ops are encountered is because that's supported via setting [pullCreds](https://opctl.io/docs/reference/opspec/op-directory/op/call/op#pullcreds). Just wanted to verify you weren't intending to change that...
Thanks, that doc link is from an example so it's a bit fast & loose. The docs about `opspec` are meant to be authoritative, heres the section on `socket` typed...
also, FWIW I think this will get rid of your need for passing in a socket alltogether https://github.com/opctl/opctl/issues/498
@apexskier thanks for filing this. This has been a long standing ask and the reason it's not been completed is lack of decision on the command interface. Your thoughts are...
@mrsadooo You can use [this op](https://github.com/opspec-pkgs/object.merge) to achieve this at the moment; the above syntax would just make it "native" rather than an op.
Yeah, any time we are adding something to the API it's gotta be piped through all layers. The implementation should be relatively straight forward though based on current architecture. At...