On Tue, 3 Sep 2024 08:54:39 GMT, Claes Redestad <redes...@openjdk.org> wrote:
>> Shaojin Wen has updated the pull request incrementally with two additional
>> commits since the last revision:
>>
>> - code style
>> - remove unsafe
>
> test/micro/org/openjdk/bench/java/lang/classfile/Utf8EntryWriteTo.java line
> 76:
>
>> 74: byte[] bytes = HexFormat.of().parseHex(
>> 75: switch (charType) {
>> 76: case "ascii" -> "78";
>
> Might be good to randomize these inputs a bit.
I also note that all Strings are the same length. With vectorization on the
table we'd want to test a healthy mix of strings of various lengths. @liach
also suggested somewhere that strings which have a leading ASCII substring then
a non-ASCII somewhere in the middle or near the end had a regression in an
earlier iteration. It would be good to have a micro test for such strings.
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/20772#discussion_r1741732133
