[foreign-memaccess+abi] RFR: 8315769: Add support for sliced allocation [v2]

[Jorn Vernee]

On Wed, 6 Sep 2023 13:33:17 GMT, Maurizio Cimadamore <mcimadamore at openjdk.org> wrote:

> No, if we do not propagate the original layout, some of the alignment checks are going to be missing.

I don't get this. I'm not saying we don't propagate the original layout, I'm just saying that the source layout should be `JAVA_BYTE` here.

The alignment checks of the destination layout only apply to the destination segment, right? Or are you saying that if the destination layout is hyper-aligned (> 1 byte), we should fail because the byte[] elements are only guaranteed to be aligned to 1 byte?

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

PR Review Comment: https://git.openjdk.org/panama-foreign/pull/878#discussion_r1317615360