[5.x] Upgrade brick/math to support versions ^0.14.2–^0.17

[balu-lt]

Description

This PR upgrades the brick/math dependency and updates the codebase to use the new rounding mode constants introduced in version 0.14.2.

The 5.x branch was targeted to keep the implementation clean and avoid legacy constraints.

Key changes:

• Breaking: Updated composer.json to require brick/math ^0.14.2 || ^0.15 || ^0.16 || ^0.17, dropping support for ^0.11–^0.13.
• Migrated all usages of Brick\Math\RoundingMode constants from SCREAMING_SNAKE_CASE (deprecated) to PascalCase (e.g., HALF_UP -> HalfUp).
• Updated CI workflow to test against the new minimum version (0.14.2).
• Adjusted BrickMathCalculatorTest to handle variations in exception messages between different brick/math versions.
• Refactored Ramsey\Uuid\Type\Time::toString() to ensure it returns a numeric-string, addressing a PHPStan static analysis error.

Motivation and context

brick/math 0.14.2 introduced PascalCase constants and deprecated the old ones. The old constants are removed in version 0.15. This update ensures the library remains compatible with current and future versions of the dependency. Additionally, it fixes a type-hinting issue reported by PHPStan in the Time class.

How has this been tested?

• Ran unit tests via vendor/bin/phpunit to ensure all functionality remains intact.
• Ran PHPStan static analysis via composer dev:analyze to verify the fix in src/Type/Time.php.
• Verified compatibility with the new minimum version of brick/math (0.14.2) in the CI environment.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

PR checklist

• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have read the CONTRIBUTING.md document.
• I have added tests to cover my changes.

[WoBBeLnl]

@ramsey really sorry to bother you but is it possible to merge this anytime soon? There are multiple PR open for the same issue, one starting february. If you need any help please let me know!

[ramsey]

Was this intended to be made to the 5.x branch? The 5.x branch does not have any releases.

[ramsey]

Was this intended to be made to the 5.x branch?

Never mind. I see why you did this. It's because you've set the minimum brick/math to 0.14.2.

Are you able to clean up the static analysis errors?

[balu-lt]

Was this intended to be made to the 5.x branch?

Never mind. I see why you did this. It's because you've set the minimum brick/math to 0.14.2.

Are you able to clean up the static analysis errors?

Yes, I targeted the 5.x branch to keep the implementation clean and avoid legacy constraints.

I’ve fixed the errors as well.

[jnoordsij]

With #638 merged I guess this is now obsolete? Or otherwise could be just a refactoring to remove the compatibility layer introduced there and solely support the newer brick/math versions.