RFR: 8360389: Support printing from C2 compiled code [v8]
Benoît Maillard
bmaillard at openjdk.org
Tue Sep 2 13:35:43 UTC 2025
On Mon, 1 Sep 2025 14:38:59 GMT, Emanuel Peter <epeter at openjdk.org> wrote:
> Ah, I see! Could you verify that somehow in your function, to catch it early? Or is there any way we could drop the need for these place-holders?
I have made modifications so that placeholder nodes are automatically inserted for `long` and `int` nodes @eme64. Thanks for the suggestion!
> FYI: I have it working now, great! 😊 The sequence of different prints can be a little messed up, but that's probably due to the controls. Just an observation, but I'm fine with that.
Yes, this is due to the controls. As suggested on slack, in the future it would be nice if the user could optionally provide a control to attach the call to instead of using the automatic procedure. This would probably help make it more predictable when needed.
> Sadly: it seems that the printing scares the bug away I'm trying to reproduce.
This will probably always be an inherent limitation unfortunately.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/26475#issuecomment-3245371745
More information about the hotspot-dev
mailing list