Not PMC, of course, but I would vote -1 on the original proposal, based on:
1. package scope is rare (unless someone proves otherwise ;-) )
2. Having a keyword and annotation is confusing for users
3. Dropping the annotation will break existing code (unless no one uses
the feature), and functionality cannot be simulated by someone who
depends on this
4. Keeping the annotation gives no benefit for IDEs (IntelliJ)
(although, after thinking about this, I feel this is a bit "The need
of the many outweigh..." in any case)
5. Giving you a little bit extra (in this case: flexibility with
regards to package scope granularity) to me typically Groovy, and
one thing that seperates it from other languages
6. In Groovy one already uses a lot of annotations in some cases
anyway, so I do not see the problem with having another one on a
class for a rare case (I do agree, however, that the frequent case
should be made as simple as possible in the annotation
implementation, if possible)
This modified idea I find better, but it still only fixes the
"functionality cannot be simulated by someone who depends on this" part
of my arguments above, and introduces the problem that the
fine-granularity feature of package scope in Groovy gets overlooked
completely, making it a dead feature.
Cheers,
mg
On 31.12.2017 05:50, Nathan Harvey wrote:
Perhaps another alternative could be introducing a @Scope annotation to apply
the scoping behavior used by @PackageScope. Eg @Scope(Modifier.PRIVATE) or
@Scope(value=PRIVATE, target=[CLASSES]).
I suppose the "package" keyword doesn't indicate the full "package private"
but I don't think that's necessary. My original point remains that using an
annotation to declare scope is awkward and cumbersome.
--
Sent from: http://groovy.329449.n5.nabble.com/Groovy-Dev-f372993.html
