Status of array covariance in lworld branch

Tobias Hartmann tobias.hartmann at oracle.com
Wed Dec 12 13:00:52 UTC 2018


On 12.12.18 08:39, David Simms wrote:
> So it's totally fine for example if C2 keeps it's rather complex code for supporting full covariance
> support, we may through a switch around it...same goes for any other previous code investments.

We will definitely keep it to avoid the overhead of removing and re-adding later (I've filed
JDK-8215283 [1] for that). It's also quite straight forward to guard it with a flag. However,
without a flag for javac/verifier/interpreter/runtime, there is no way we can test this code and
prevent it from rotting..

Best regards,
Tobias

[1] https://bugs.openjdk.java.net/browse/JDK-8215283


More information about the valhalla-dev mailing list