class constant in homebrew formula files can that be a thing?

[ipatch]

i recently updated a formula in the homebrew-freecad tap to use a "class constant" to make updating / versioning the formula a little easier for me, but unfortunately i'm now getting these errors when attempting to run brew style on the freecad repo/tap. ie.

╰─λ brew style freecad/freecad
/home/capin/homebrew/Cellar/ruby/4.0.1/bin/bundle
An error occurred while FormulaAudit/DependencyOrder cop was inspecting /home/capin/homebrew/Library/Taps/freecad/homebrew-freecad/Formula/freecad@1.0.2_py313_qt6.rb:4:0.
To see the complete backtrace run rubocop -d.
Inspecting 57 files
....................An error occurred while FormulaAudit/DependencyOrder cop was inspecting /home/capin/homebrew/Library/Taps/freecad/homebrew-freecad/Formula/freecad@1.0.2_py313_qt6.rb:4:0.
To see the complete backtrace run rubocop -d.
.....................................

57 files inspected, no offenses detected

1 error occurred:
An error occurred while FormulaAudit/DependencyOrder cop was inspecting /home/capin/homebrew/Library/Taps/freecad/homebrew-freecad/Formula/freecad@1.0.2_py313_qt6.rb:4:0.
Errors are usually caused by RuboCop bugs.
Please, update to the latest RuboCop version if not already in use, and report a bug if the issue still occurs on this version.

Mention the following information in the issue report:
1.85.1 (using Parser 3.3.10.2, rubocop-ast 1.49.0, analyzing as Ruby 2.7, running on ruby 4.0.1) [aarch64-linux]

the lines in question that i think causes this issue,

PY_VER = "3.13".freeze

depends_on "python@#{PY_VER}"

https://github.com/FreeCAD/homebrew-freecad/pull/798/changes

and sure enough changing the below line back resolves the style errors,

 depends_on "python@3.13"

[p-linnane]

We define a method for this in formulae:

https://github.com/Homebrew/homebrew-core/blob/2f71efb7cc63350e94fdc9f356656e57d4031170/Formula/c/cython.rb#L25-L27

[ipatch]

yeah but that's not a class constant, i was hoping to reference PY_VER throughout the entire file. however, it's not a deal-breaker to have to update the single reference in the depends_on but i just thought i'd bring it up.

[Bo98]

This seems like a bug with our FormulaAudit/DependencyOrder cop. While we can't properly enforce an order when you use dynamic strings like that, we should definitely not be erroring like it currently is.

Homebrew/brew#21750