Stephan Renatus
Stephan Renatus
> Of all the verbs booted from the list, I think %T (prints type-of-value) might be worth keeping for debugging purposes alone. It doesn't need to happen right away, but...
Looks like `%T` isn't as bad as I thought -- we're not leaking golang struct types -- but it's still got some things to fix: call | output | expected...
> We could remove %x/%X entirely. (Removes the problem case, but also a very useful formatting option for bitwise operations. 😞 ) Those bitwise ops would be using integers, only,...
```rego x := sprintf("%x", [100]) y := sprintf("%x", [100.0]) ``` ⬇️ ``` { "x": "64", "y": "0x1.9p+06" } ``` I'm an ignorant, I have no idea what `y` means. 😅...
Haven't yet looked closely. Does it also happen if we're iterating a sub-strucutre, like `data.foo[x]`?
It looks like we're hitting this condition where: 1. `doc` contains your data, `{"a": {"metadata": "interesting"}, "b": {"metadata": "other" } }` [here](https://github.com/open-policy-agent/opa/blob/v0.41.0/topdown/eval.go#L2048) and 2. `e.node`, the retrieved rule tree node,...
Fun fact, wasm does the right thing: ``` % opa eval -t wasm -fpretty -d scratch/duplicate-keys/p.rego -d scratch/duplicate-keys/d.json data.a.example1 [ "a", "b" ] ```
💭 @jeniawhite @eyalkraft @uri-weisman Perhaps you would like to give it a shot? 😃
Hello from the OPA side 👋 > [...] there doesn't appear to be a way to distribute something to using OPA users that doesn't require creating a new build of...
> has OPA considered adding plugins that could be useful to the community (such as the SpiceDB one) to the main build? Here's the list of built-in functions currently supported:...