Request: implement/add ExtensionRegistry.isEmpty() in Java

[protobufel]

It is crucial to know whether an ExtensionRegistry is empty - for a number of reasons; but impossible at the moment. It could have been produced as ExtensionRegistry.getEmptyRegistry(), or as the result of ExtensionRegistry.newInstance(), or ExtensionRegistry.newInstance().getUnmodifiable(), and so forth.
However, there is no access to the underlying map, or its length, and also there is no equals.

[xfxyjwf]

Can you expand on the reasons to check whether an ExtensionRegistry is empty?

[protobufel]

Use case, one of many: you have a nonnull field, parameter of type
ExtensionRegistry, empty registry would be its null-token. How then you
recognize it later? The registry could be also empty as described earlier.

Or you've got a registry value, and if it is empty you do nothing,
otherwise some processing, for example, reparsing/deserialization.

The feature is safe, and doesn't break anything!
On Oct 14, 2014 7:16 PM, "Feng Xiao" [email protected] wrote:

Can you expand on the reasons to check whether an ExtensionRegistry is
empty?

—
Reply to this email directly or view it on GitHub
#52 (comment).

[haberman]

This feature request is very old. I don't think in general we attempt to give every type an empty/null value that is recognizable. If you want a value to be possibly absent, it seems to make more sense to just make it nullable, instead of making it nonnull but having an "empty" value.