[
https://issues.apache.org/jira/browse/CASSANDRA-14716?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16700148#comment-16700148
]
Sam Tunnicliffe commented on CASSANDRA-14716:
---------------------------------------------
{quote}Is it too late to change {{Adler32}} to {{ADLER32}} at this point? Would
it be desirable?{quote}
No, I think that's acceptable and desirable, if only for consistency. This is
the only usage of {{ChecksumType}} where the algo is configurable, every other
usage is hardcoded to {{CRC32}}. We included support for {{Adler32}} in
CASSANDRA-13304, rather than just removing that {{ChecksumType}} to enable
future potential perf improvements to be taken advantage of
(https://issues.apache.org/jira/browse/CASSANDRA-13304?focusedCommentId=15923191&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15923191).
Before committing I'll rename the enum value, re-run CI and post a message to
dev list to be sure nobody objects to this post-freeze change. Thanks.
> Protocol frame checksumming options should not be case sensitive
> ----------------------------------------------------------------
>
> Key: CASSANDRA-14716
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14716
> Project: Cassandra
> Issue Type: Bug
> Components: CQL
> Reporter: Sam Tunnicliffe
> Assignee: Sam Tunnicliffe
> Priority: Major
> Fix For: 4.0
>
>
> Protocol v5 adds support for checksumming of native protocol frame bodies.
> The checksum type is negotiated per-connection via the \{{STARTUP}} message,
> with two types currently supported, Adler32 and CRC32. The mapping of the
> startup option value requested by the client to a \{{ChecksumType}} should
> not be case sensitive, but currently it is.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
