On Wed, 21 Jun 2023 06:50:20 GMT, Chen Liang <li...@openjdk.org> wrote:

>> Adam Sotona has updated the pull request incrementally with one additional 
>> commit since the last revision:
>> 
>>   ClassfileBenchmark::transformWithNewMaps changed to transformWithAddedNOP
>
> src/java.base/share/classes/jdk/internal/classfile/impl/ClassfileImpl.java 
> line 63:
> 
>> 61:             DeadCodeOption.PATCH_DEAD_CODE,
>> 62:             DeadLabelsOption.FAIL_ON_DEAD_LABELS,
>> 63:             new 
>> ClassHierarchyResolverOptionImpl(ClassHierarchyResolver.defaultResolver()),
> 
> Don't users of the default context still share a non-thread-safe-cached 
> version of the default resolver? We might still need to convert 
> DEFAULT_CONTEXT field to a static defaultContext() method that returns a `new 
> ClassfileImpl(...);` (the attribute mapper option can be shared).

Please disregard this comment: Now the distinction of cache of resolvers is 
done in `StackMapGenerator`, meaning each stack map generation task has its own 
cache. Didn't see it early on.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/14180#discussion_r1236465502

Reply via email to