RFR: 8275338: Add JFR events for notable serialization situations [v3]

Raffaello Giulietti rgiulietti at openjdk.org
Tue Dec 19 17:44:52 UTC 2023


On Tue, 19 Dec 2023 17:13:58 GMT, Erik Gahlin <egahlin at openjdk.org> wrote:

>> The intent is that they are stable and for programmatic usage, whereas the message is more for human consumption. The codes are used in the test, for example, and are declared as public static in the event classes.
>> 
>> Alternatively, one could parse the message, but that's less robust in face of changes, I think.
>
> Users (not OpenJDK developers) don't know what the error code means. I think it's better to not have them. This is how other events work. If you want to guard against changes, I would export the package to the test.

What about fixed `String`s rather than `int`s for the kind of error?
Something like `"SUID_INEFFECTIVE_ON_ENUM"`, and so on?
It would be nice to be able to use enums, but AFAIK that's not supported in JFR.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17129#discussion_r1431744479



More information about the security-dev mailing list