Today, the JFR parser validates that IDs in a checkpoint event are valid, meaning that there is a corresponding object written for the ID that is found. This validation is done when running with the -esa option.
Unfortunately, this is not sufficient to detect all missing IDs.
- The parser caches values for IDs that are read in earlier chunks, which is an optimization to improve performance. However, this can lead to the parser finding an object even though it is not in the current chunk.
- ID values that are written as event values are not validated.
These new assertions cannot be checked in until the underlying issues have been fixed.
Unfortunately, this is not sufficient to detect all missing IDs.
- The parser caches values for IDs that are read in earlier chunks, which is an optimization to improve performance. However, this can lead to the parser finding an object even though it is not in the current chunk.
- ID values that are written as event values are not validated.
These new assertions cannot be checked in until the underlying issues have been fixed.
- relates to
-
JDK-8353227 JFR: Prepare tests for strong parser validation
-
- In Progress
-
-
JDK-8353226 JFR: emit old object samples must be transitive closure complete for segment
-
- In Progress
-