On Wed, 27 Nov 2024 16:27:34 GMT, Per Minborg <pminb...@openjdk.org> wrote:
> This PR proposes to check for empty layouts in the fallback linker and then
> prevent them from being forwarded.
>
> Passes tier5 linker tests.
Marked as reviewed by jvernee (Reviewer).
test/jdk/java/foreign/TestLinker.java line 254:
> 252: var struct8a8 = MemoryLayout.structLayout(sequence0a8,
> sequence3a1, padding5a1);
> 253: var fd = FunctionDescriptor.of(struct8a8, struct8a8, struct8a8);
> 254: if (linker.getClass().equals(FallbackLinker.class)) {
Alternatively, we could move this test case to a separate file (perhaps with
other test cases that test empty aggregates), and then exclude it for the
fallback linker using `@requires jdk.foreign.linker != "FALLBACK"`.
I don't mind this approach though.
-------------
PR Review: https://git.openjdk.org/jdk/pull/22417#pullrequestreview-2467939114
PR Review Comment: https://git.openjdk.org/jdk/pull/22417#discussion_r1862058562
