RFR: 7903613: Bad nested names are sometimes attached to structs [v10]
Maurizio Cimadamore
mcimadamore at openjdk.org
Thu Jan 11 16:10:57 UTC 2024
On Thu, 11 Jan 2024 14:18:28 GMT, Jorn Vernee <jvernee at openjdk.org> wrote:
>
> Are we sure we're filtering out the right duplicates? e.g. if we filter out the N + 1 definition, doesn't the actual struct class we generate appear at the wrong level?
My understanding here is that it doesn't matter much _which_ we retain, as long as (a) we have a consistent strategy to generate struct names based on what we keep (e.g. whether toplevel or "nested"), and (b) we don't end up generating things twice, redundantly. As a result, picking the "wrong" one can only result in a suboptimal name being generated (at least that's the theory).
-------------
PR Comment: https://git.openjdk.org/jextract/pull/167#issuecomment-1887490543
More information about the jextract-dev
mailing list