[
https://issues.apache.org/jira/browse/CASSANDRA-1196?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12884047#action_12884047
]
Folke Behrens commented on CASSANDRA-1196:
------------------------------------------
That unit test only works because Java prepends the byte order mark to the
output. If that is really what you wanted you could have initialized the byte
array to 0xFF, 0xFE. Looks a lot less coincidental.
> Invalid UTF-8 data should cause exceptions
> ------------------------------------------
>
> Key: CASSANDRA-1196
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1196
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Stu Hood
> Assignee: Nick Bailey
> Priority: Minor
> Fix For: 0.7
>
> Attachments: 0001-Initial-fixes-to-utf-decoding.patch,
> 0002-Unit-test-for-UTF8-fixes.patch
>
>
> Our current method for decoding UTF-8 data in OrderPreservingPartitioner and
> CollatingOrderPreservingPartitioner will silently decode invalid UTF-8 data.
> This may also be a problem UTF8Type.
> Instead, we should probably throw an exception, since bad UTF-8 data means
> either user error or corruption.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
