RFR [8014657] CheckedInputStream.skip allocates temporary buffer on every call (original) (raw)

Ivan Gerasimov ivan.gerasimov at oracle.com
Thu May 16 11🔞55 UTC 2013


On 16.05.2013 13:04, Alan Bateman wrote:

On 15/05/2013 22:19, Ivan Gerasimov wrote:

:

Of course you're right! I should have waited for the compilation to finish before posting the message. Here's the updated webrev: http://cr.openjdk.java.net/~dmeetry/8014657/webrev.1/ <http://cr.openjdk.java.net/%7Edmeetry/8014657/webrev.1/> The updated webrev looks okay to me. So have you done much testing with this and do you have any performance data to share? Thanks for the review!

No, I haven't done any testing at all (yet.) I'm not sure what can be tested here - is it performance improvement of GC in the case of lots of calls to the skip function?

What kind of testing would you suggest?

I relied mostly on two things - assumption that one allocation is better than several and the fact that the other input streams use the same approach with reused temporary buffer.

Sincerely, Ivan

-Alan



More information about the core-libs-dev mailing list