[
https://issues.apache.org/jira/browse/CASSANDRA-9499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14587997#comment-14587997
]
Sylvain Lebresne commented on CASSANDRA-9499:
---------------------------------------------
bq. the existing scheme (including a positive-only variant supporting 250
values)
Again, I haven't really looked at the existing scheme so I really wasn't fixed
on it. I was, in fact, more thinking of the simpler scheme of using one bit as
a "continuation bit", which does only support 127 values in a single byte but
degrades to 2 bytes properly. Anyway, doesn't matter.
bq. the worse our instruction cache population becomes
I'm unfortunately not smart enough to properly estimate how that kind of effect
would play against other considerations (like saving bytes sometimes) on
average in the grand scheme of everything. My point was merely to suggest that
expressing our assumptions through different API calls that would be optimized
based on these assumption could be an idea worth exploring, but I don't want to
push it any more than that. I also don't have strongly objections to the scheme
you're suggesting.
> 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)
