RFR: 8343306: javac is failing to determine disjoint-ness between a class and a sealed interface [v3]

[Maurizio Cimadamore]

On Thu, 31 Oct 2024 17:57:45 GMT, Vicente Romero <vromero at openjdk.org> wrote:

>> For code like:
>> 
>> 
>> class Test {
>>     sealed interface I permits C1 {}
>>     non-sealed class C1 implements I {}
>>     class C2 extends C1 {}
>>     class C3 {}
>>     I m(int s, C3 c3) {
>>         I i = (I)c3;
>>     }
>> }
>> 
>> javac is failing to issue an error and accepts this code. The spec is clear stating that code like this should be rejected. See:
>> 
>> 5.1.6.1 Allowed Narrowing Reference Conversion:
>> 
>> • A class named C is disjoint from an interface named I if (i) it is not the case that
>>   C <: I , and (ii) one of the following cases applies:
>>   – C is freely extensible (§8.1.1.2), and I is sealed , and C is disjoint from all of
>>   the permitted direct subclasses and subinterfaces of I .
>> 
>> and just below it continues:
>> • A class named C is disjoint from another class named D if (i) it is not the case
>> that C <: D , and (ii) it is not the case that D <: C .
>> 
>> so here we have the `C3` is a freely extensible class and interface `I` is sealed and `C3` is disjoint from `C1` which is is the permitted subclass of interface `I`
>> 
>> This PR should sync javac with the spec
>> 
>> TIA
>
> Vicente Romero has updated the pull request incrementally with one additional commit since the last revision:
> 
>   implementation closer to the spec

The new code looks great. I pulled up the new file and read the code side by side with the spec and was very pleased with the easy-to-spot symmetry. Thanks!

src/jdk.compiler/share/classes/com/sun/tools/javac/code/Types.java line 1689:

> 1687:                     return false;
> 1688: 
> 1689:                 if (ts.isInterface() && !ss.isInterface() || // case I: one is a class and the other one is an interface

Nit: you can still simplify this to `ts.isInterface() != ss.isInterface()`

-------------

Marked as reviewed by mcimadamore (Reviewer).

PR Review: https://git.openjdk.org/jdk/pull/21794#pullrequestreview-2408852261
PR Review Comment: https://git.openjdk.org/jdk/pull/21794#discussion_r1825025286