chore: sketch of the power-line concept #3
Conversation
Signed-off-by: Irakli Gozalishvili <contact@gozala.io>
Signed-off-by: Irakli Gozalishvili <contact@gozala.io>
Signed-off-by: Irakli Gozalishvili <contact@gozala.io>
Signed-off-by: Irakli Gozalishvili <contact@gozala.io>
Signed-off-by: Irakli Gozalishvili <contact@gozala.io>
Signed-off-by: Irakli Gozalishvili <contact@gozala.io>
notes/conds.md
Outdated
| ### concrete proposal for the inital release | ||
|
|
||
| Here's a sketch that shows the features | ||
|
|
||
| ```js | ||
| { | ||
| "iss": "did:key:alice", | ||
| // ... | ||
| "policy": [ | ||
| ["var", ["?x", "?y"], // The `?` here is just by convention | ||
| ["args", "foo..bar.[].baz", "?x"], | ||
| [["some", "match"], "?x", "*@example.com"], | ||
|
|
||
| ["args", "foo.quux", "?y"], | ||
| [">", "?y", 42], | ||
| ["<", "?x", "?y"] | ||
| ] | ||
| ] | ||
| } | ||
| ``` | ||
|
|
||
| Of note: | ||
|
|
||
| * Selector syntax based on JSONPath / JSON Poiniter / jQ | ||
| * `"foo..bar.[].baz"` ~ `foo.{*recursive descent*}.bar.{ALL}.baz` | ||
| * `var` is like `fresh` in minikanren: introduces logic variables | ||
| * these give the ability to run predicate logic in the middle of a sector | ||
| * `[[<node>, <path>, "?x"], [<pred>, "?x", <expr>]` | ||
| * Declarative matching from `args` onwards | ||
| * Anything of the form `[<node>, <selector>, <value or variable>]` | ||
| * Can avoid namespace conflicts because users define variable names locally | ||
| * Recurses if needed | ||
| * A trivial case: `[["args", "some.path", "?x"], ["?x", "more.path", 42]]` | ||
| * Matches `args.some.path.more.path == 42` | ||
| * Use of `args` to start the path allows us to open up the synatx over time if we want | ||
| * Plus it's very declarative |
There was a problem hiding this comment.
Perhaps worth noting that the above also lets us avoid == in a lot of common cases, too:
["args", "some.path", 42] is like {==: {args: {some: {path: 42}}}}
There was a problem hiding this comment.
Odd thing about ["args", "some.path", "?x"] is that it's bit unclear if e.g. this would be valid
["", "args.some.path", "?x"] or this ["args.some", "path", "?x"]
I feel like it may make more sense to think of args as a built-in pre-poulated variable. perhaps ? could be that. Or maybe for built-in variables we could use different prefix so we could introduce new ones in the future without colliding with user defined variables.
There was a problem hiding this comment.
["", "args.some.path", "?x"]
I think that we can't run the selector on "", so it should fail, right?
perhaps ? could be that
Yeah, exatcly: I was thinking of args as that pre-populated variable. How about ?args for consistent variable syntax?
There was a problem hiding this comment.
so we could introduce new ones in the future without colliding with user defined variables.
Not a bad idea, though I think it's up to the user to avoid collisions. We have a schema type in the capsule, so we know what's reserved.
We could introduce more syntax like $env ($args), which feels maybe appropriate? It's is from the $environment
Co-authored-by: Irakli Gozalishvili <contact@gozala.io> Signed-off-by: Brooklyn Zelenka <be.zelenka@gmail.com>
Co-authored-by: Irakli Gozalishvili <contact@gozala.io> Signed-off-by: Brooklyn Zelenka <be.zelenka@gmail.com>
Co-authored-by: Vasco Santos <santos.vasco10@gmail.com> Signed-off-by: Irakli Gozalishvili <contact@gozala.io>
Breaking out some of the discussion points into separate issues
some#7everyexpression #8