[
https://issues.apache.org/jira/browse/CASSANDRA-9499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14593474#comment-14593474
]
Benedict commented on CASSANDRA-9499:
-------------------------------------
Looking good. I've pushed some suggestions to the bit fiddling, for both
clarity and performance
[here|https://github.com/belliottsmith/cassandra/tree/C-9499-suggestions].
Including efficient size computation, some extraneous operation removal, and
what seemed to me clearer names to methods (please feel free to dispute). I had
to strip out the debugging statements to figure out what was going on, hope
that isn't a problem for you.
It's still more computationally involved than I would _like_, and I feel there
may be some more efficiencies. So I will mull it over a little more, but wanted
to upload what I had so we didn't tread too much on each other.
Some further suggestions I did not implement:
* I'm not sure why we need the try/finally block, and I'd prefer we removed it
- there's nothing that should be able to throw an exception between our setting
and unsetting the limit.
* The extra reading in NIODataInputStream, I would prefer we extract to a
separate method, and force it to not be inlined. Since it should be a rare
occurrence that we need to do this, and we incur more expensive costs than this
inner method invocation when this does happen, I would prefer to keep the rest
of the method as tightly packed for icache behaviour as possible
> Introduce writeVInt method to DataOutputStreamPlus
> --------------------------------------------------
>
> Key: CASSANDRA-9499
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9499
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: Benedict
> Assignee: Ariel Weisberg
> Priority: Minor
> Fix For: 3.0 beta 1
>
>
> CASSANDRA-8099 really could do with a writeVInt method, for both fixing
> CASSANDRA-9498 but also efficiently encoding timestamp/deletion deltas. It
> should be possible to make an especially efficient implementation against
> BufferedDataOutputStreamPlus.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
