Null as Identifier in HCL

[MBoegers]

What's changed?

null is a valid key for objects and variables in HCL. We blocked it as syntactically error.

What's your motivation?

allow basic setups as

terraform {
  required_providers {
    null = {
      source = "hashicorp/null"
      version = "=3.2.3" 
    }
  }
}

Anything in particular you'd like reviewers to focus on?

Anyone you would like to review specifically?

Have you considered any alternatives or workarounds?

null seems to be common in TF

Any additional context

Checklist

• I've added unit tests to cover both positive and negative cases
• I've read and applied the recipe conventions and best practices
• I've used the IntelliJ IDEA auto-formatter on affected files

[MBoegers]

I know we usually don't force push, I just neede to get the PR displaying right

[timtebeek]

Somewhat surprised not to see any corresponding change in the tree package here around:

rewrite/rewrite-hcl/src/main/java/org/openrewrite/hcl/tree/Hcl.java

Line 65 in 05795eb

class Attribute implements BodyContent, Expression {

Is that just a problem with my expectations? Could you help explain why such a change wasn't needed?

[MBoegers]

Token NULL here is not the OOP null as no object. In the G4 the NULL is a special identifier with the value "null" and will be parsed into a Hcl.Identifier with the name="null".

The NULL-Token is only used in 2 locations to selectively allow "null" as an identifier in these locations (as object keys and variable names). This way we still reject corrupt files at the parser level when null is used in a different location or no identifier is provided.

You see no change in the Hcl.Tree because none is needed to represent an Hcl.Identifier with value "null". And I choose the minimal change.
We could think about mirroring this setup with a specialization of Hcl.Identifier to explicitly model "null" as an identifier in these cases.
I think the gain would be very low, as null is simply valid in these cases, and the risk of introducing new LST elements is higher.

For example:

terraform {
required_providers {
  null = {
    source = "hashicorp/null"
    version = "=3.2.3"
  }
}

is parsed into (sorry for no TreeVisitingPrinter):

[MBoegers]

While writing these lines, I realize that NULL is probably not a great name for the Token.

[greg-at-moderne]

The g4 change and related changes look fine to me.

I am only slightly concerned with the changes to the JsonPath* classes. It's hard to tell for sure, but I hope they are only whitespace change as there's no reason for these to be altered. Github's "Hide Whitespace" option in the review interface hides only some of them, but they all look trivial.

[MBoegers]

agree they look strange let me reset to main

[MBoegers]

@greg-at-moderne the answer is, I executd the ./gradlew rewrite-hcl:generateAntlrSources task which also generated the HCL internal JSON grammar. rolled back because its a disturbing change