[PHP 8.1] Add basic support for Spatie Enums to native Enums

[mallardduck]

Per the title, this add support for the most basic usage of Spatie Enums.

This does not cover usages of Spatie Enums where a values() method is use to override the automatic values. However it does (from my understanding) mimic the experience one would expect with only using the DocBlocks to define enum cases for their enums.

It also doesn't bother to remove the import of the spatie Enums since that seems to match what the MyCLabsClass version of the rector.

[Gummibeer]

Awesome! 🥳
Thanks for taking care Dan! 🙏

[TomasVotruba]

Thank you for nice and clean code. It's joy to read ☺️
One little detail and it's ready to go

[samsonasik]

I think the Enum suffix should be kept to avoid manual file rename.

[mallardduck]

@samsonasik - I suppose that's a fair view of things.
People will just have to de-suffix and rename it manually.

Well unless there's a way to have rector rename the files too. Then maybe we just do that. 🤔

[mallardduck]

@TomasVotruba - Am I mistaken or would the UpdateFileNameByClassNameFileSystemRector rector potentially cover correcting the filename to match? I'm guessing it depends on if the enum is considered a ClassLike node?

[samsonasik]

The EnumFactory used by other rule: MyCLabsClassToEnumRector which kept the Enum suffix, the functionality should be same for consistency.

You may need to use Renaming and Restoration rules to rename class and file.

[mallardduck]

@samsonasik - I disagree that this rector needs to match how MyCLabs rector works. These are both distinct and opinionated packages providing Enum features before native Enums. As such I think matching the individual packages convention will make sense.

One of the creators/maintainers of the Spatie package @brendt wrote an article about how they will be updating their Spatie Enums to native Enums. And that included the removal of redundant *Enum suffix on the enum - makes perfect sense since we don't suffix our classes as *Class.

EDIT: Link to Article: https://stitcher.io/blog/php-enums

[mallardduck]

I guess put a different way, what's the Rector philosophy on providing rectors of varying levels of aggressiveness?

Personally I want a rule that will update all the things for me and I don't like redundant suffixes. So having one rector that gets me part the way there doesn't feel like a solution to me. At that point I'd just do this all manually since it will still require me to rename the enum and file.

[samsonasik]

I think rename class name is not this rule's responsibility, you can use Renaming and Restoration rules for renaming class and file used along with this rule.

[TomasVotruba]

I think rename class name is not this rule's responsibility, you can use Renaming and Restoration rules for renaming class and file used along with this rule.

👍 Agreed, the rule should be doing one job at a time.

Let's merge this as it is to keep it simple and iteratate improvements 🙂 Thank you

[mallardduck]

the rule should be doing one job at a time.

Fair enough - that's why I was seeking to understand the philosophy here.

[samsonasik]

Thank you @mallardduck

[TomasVotruba]

Fair enough - that's why I was seeking to understand the philosophy here.

Undrestood. We try to follow unix philosophy. The oldest rules try to do many things at once, but last 2 years we try to go as narrow as possible, while keeping the main idea together.