Support @properties that are 'copy' in addition to the default 'retain' about mogenerator HOT 12 CLOSED

I'm down with that, implementing it the way your described. I'd accept such a patch if you want to work it up. :-)

from mogenerator.

Cool. I could try, but won't have time for a few weeks. I wanted to at least get the ball rolling with this bug entry.

So how about key 'attributeValueSetterSemantics' where the value can be 'retain' or 'copy', and if there's no value it defaults to retain?

from mogenerator.

I know about the busy thing :-) attributeValueSetterSemantics sounds fine.

from mogenerator.

Sounds like #84 supersedes this issue? Should I close it?

from mogenerator.

I don't think they're the same, but certainly related.

Regardless of memory model, there is a difference between copying and referencing an object.

It could still be useful, IMHO, it be able to tag an attribute as either 'copy' or 'retain aka strong'.

from mogenerator.

Here's a first shot: https://github.com/diederich/mogenerator/tree/propertiesWithCopy
Instead of allowing to define retain/copy in the userInfo dictionary, this approach uses a boolean value. I'm not sure if this is a good approach, but it works for the above example, as "strong" implies copy. Though others like weak, __unsafe_unretained might be interesting for the PONSO templates.

My current use case for this is an NSManagedObject which should implement a protocol that defines a property with copy semantics.

from mogenerator.

What do the two states of your boolean represent? What do you mean "strong implies copy"?

from mogenerator.

The state says either to use copy semantics or not.
If it's set to yes, the property is declared as

@property (nonatomic, copy) <Type> <Name>;

if set to no it's dependent on ARC. If ARC is enabled it's

@property (nonatomic, strong) <Type> <Name>;

without ARC it's

@property (nonatomic, retain) <Type> <Name>;

So there's no special case for ARC and copy. If you declare a property as with copy semantics, strong is implicitly added (http://clang.llvm.org/docs/AutomaticReferenceCounting.html#ownership.spelling.property)

from mogenerator.

I think you meant to say "copy implies strong" not "strong implies copy".

from mogenerator.

Definitely, thanks.

from mogenerator.

related: #84

from mogenerator.

This is ancient. Closing it out. If its still an issue for you, please open a new issue.

from mogenerator.