Task #27627: Overhaul annotation support/syntax - TYPO3.Flow - TYPO3 Forge

TYPO3 Flow Base Distribution TYPO3.Eel TYPO3.Flow TYPO3.Fluid TYPO3.Kickstart TYPO3.Welcome Packages Applications

Task #27627

Overhaul annotation support/syntax

Added by Karsten Dambekalns about 4 years ago. Updated almost 4 years ago.

Status:

Resolved

Start date:

2011-06-21

Priority:

Must have

Due date:

Assigned To:

Karsten Dambekalns

% Done:

100%

Category:

Reflection

Target version:

TYPO3 Flow Base Distribution - 1.0.0

Sprint:

Has patch:

PHP Version:

Complexity:

Description

I think it would be wise to decide on this before 1.0.0, the earlier the better. It has been discussed here and there in the past, but we need to get this thought through.

So, here are the basic questions:

Do we adopt UpperCamelCase and case-sensitivity, like Doctrine 2 and Symfony 2 use it?
Do we switch annotation parsing to the Doctrine Common library?

The case thing means

consistency with Doctrine annotations we have in place already
inconsistency with "normal" docblock API annotations
the latter can be softened by using prefixes on annotations and have them lowercased, see http://symfony.com/blog/symfony2-annotations-gets-better for examples

The parser library switch would

move the burden of maintaining a parser away from us
force us to adapt the current annotation syntax
possibly allow more flexibility

Related issues

History

#1 Updated by Bastian Waidelich about 4 years ago

I would go for 1 & 2.
If I got it right, this would change the annotations like this:

@scope prototype -> @Scope(type="prototype")
@scope singleton -> @Scope(type="singleton")
@validate StringLength(minimum = 1, maximum = 100) -> @Validate(validator="StringLength", options="{"minimum" = "1", "maximum" = "100"}")

I don't think, that the inconsistency with "normal" docblock API annotations is a problem, because IMO there's a "mental gap" between those anyways, but that's just my subjective opinion.

#2 Updated by Sebastian Kurfuerst about 4 years ago

1): +1 from my side
2): it definitely makes sense to use another framework, as maintaining such a parser always sucks ;)

If I got it correctly, we'd need to decide between the following annotation formats:

@Scope(type="prototype") vs @f3:Scope(type="prototype")
@Validate(validator) vs @f3:Validate....

So we still have to decide whether we put all FLOW3 annotations into F3\FLOW3\Annotations (so that they can be referenced without namespace), or if we define prefixes. For validation, things like: