RFR' [performance] Avoid of usage ASM.Type from InnerClassLambdaMetafactory
aleksey.shipilev at oracle.com
Wed Sep 11 10:33:29 PDT 2013
On 09/11/2013 08:23 PM, Sergey Kuksenko wrote:
This looks like a promising change, which does seem to be equivalent to
the current code. I have only the stylistic comments:
- implementation notes missing for new methods; I understand these are
right from bytecode spec, but it seems a good idea to put the
excerpts/explanations for otherwise magic numbers in there.
- the named locals in the current code are very telling; can you keep
from inlining them where possible? Does it hurt interpreter performance
to keep locals like in the current code?
More information about the lambda-dev