RFR: 8350864: C2: verify structural invariants of the Ideal graph [v5]

[Marc Chevalier]

On Tue, 9 Sep 2025 08:22:41 GMT, Emanuel Peter <epeter at openjdk.org> wrote:

>> Can you add a comment, why it can be arbitrarily large?
>> Do you have an example where we have very many ctrl uses?
>
> Also: are these all supposed to be projections of a specific kind? We could also test for that. You can also add that to a future RFE.

> Can you add a comment, why it can be arbitrarily large?

Maybe I'm very wrong about what is a CatchNode, but:

try {
...
}
catch ( ... ) { ... }
catch ( ... ) { ... }
catch ( ... ) { ... }

4 outputs: 3 handlers + 1 fallthrough.

>  Also: are these all supposed to be projections of a specific kind? We could also test for that. You can also add that to a future RFE.

I'd rather do it separately. We can always check more things, but we need to draw the line and that is safe to add later.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/26362#discussion_r2333873366