On Wed, 8 Jun 2022 14:44:33 GMT, Jim Laskey <jlas...@openjdk.org> wrote:

>> To take optimal advantage of the pre-existing optimization for repeated 
>> filters we could split the application of different types of stringifiers.
>> The resulting difference in order of evaluation is not observable by 
>> conventional means since all reference type share the same object 
>> stringifier, and the others are filtering primitives (floats and doubles) 
>> which have been passed by value already. 
>> This change neutral on many concatenation expression shapes, but for any 
>> complex expressions with interleaving float/double and reference parameters 
>> it brings a straightforward reduction in rebinds and underlying LFs 
>> generated. For example on the 
>> [MixedStringCombinations.java](https://gist.github.com/cl4es/08fb581dece3a73e89bfa52337bc4248)
>>  test there's a modest 2% reduction in total classes loaded with this change 
>> (from 16209 to 15872)
> I make heavy use of the `mh = filterArgument(baseMH, 0, arrayOfFilters);` 
> pattern. I assumed it generated fewer LF than making individual calls. Am I 
> wrong?

@JimLaskey no, you're generally right to assume that bunching them up reduces 
the number of LFs created. There's a flaw in the `makeRepeatedFilterForm` 
transform that means the theoretical total amount of forms can surpass the 
baseline if you do _both_ one-by-one and all-at-once transforms from the same 
root MH. That's a very big if, but it's something I've felt slight unease about 
since I did that optimization.


PR: https://git.openjdk.java.net/jdk/pull/9082

Reply via email to