XenoAmess commented on pull request #118:
URL: https://github.com/apache/commons-io/pull/118#issuecomment-755419401


   > Hello @XenoAmess
   > We already have a lot of changes for the next release, so I want to manage 
expectations such that I would I prefer to get out 3.12 before making even more 
big changes like these.
   > 
   > But still, let's continue this thread. Starting with the lowest-level 
bits: we need to justify the addition of the misnamed `Unsync*` classes, the 
prefix should be `Unsynchronized` like our existing 
`UnsynchronizedByteArrayInputStream`, which I've already mentioned. Perhaps our 
addition of UnsynchronizedByteArrayInputStream was a mistake since the Java 
folks make it sounds like these are superfluous in 
https://bugs.openjdk.java.net/browse/JDK-4097272, so I think we need to see a 
performance test that shows there is a clear performance benefit to adding 
those as valuable on their own.
   > 
   > We need performances test that show the differences, if any, between the 
JRE's classes and our proposed `Unsynchronized` versions. Since you propose two 
such classes `UnsyncBufferedInputStream` and `UnsyncBufferedReader`, that's two 
new tests. Or did I miss these here?
   > 
   > I think you should create a new PR for just these two new classes and 
their tests. This will make the work simpler for everyone when reviewing and 
testing.
   > 
   > TY.
   
   @garydgregory Hi. I done the performance test at 
https://github.com/apache/commons-io/pull/184.
   The two classes is faster, but not as that faster as we thought.
   Detailed performance tests results: https://pastebin.ubuntu.com/p/TtDxYw4WVG/


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to