[
https://issues.apache.org/jira/browse/ARTEMIS-2059?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16595028#comment-16595028
]
ASF GitHub Bot commented on ARTEMIS-2059:
-----------------------------------------
Github user franz1981 commented on the issue:
https://github.com/apache/activemq-artemis/pull/2274
@tabish121 @gemmellr I will be probably able to take a further look on it
after wednesday...
Anyway I think that this PR as it is will just fix the bug at a performance
cost: ByteBufUtil.utf8Bytes has a O(n) cost while ByteBufUtil.utf8MaxBytes is
O(1).
A trivial fix with no performance hit that assume that the ByteBuffer has
been sized with enough space to hold the encoded string would be to use
```ByteBufUtil.reserveAndWriteUtf8(nettyBuffer, value,
nettyBuffer.writableBytes());``` instead, but I'm not sure if
NettyWritable::put(String) can safely assume it.
> NettyWritable should use UTF-8 exact length to encode strings
> -------------------------------------------------------------
>
> Key: ARTEMIS-2059
> URL: https://issues.apache.org/jira/browse/ARTEMIS-2059
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Components: AMQP
> Reporter: Francesco Nigro
> Assignee: Francesco Nigro
> Priority: Minor
> Fix For: 2.7.0
>
>
> NettyWritable.put(String) tries to enlarge the buffer used to write a UTF-8
> string until ByteBufUtil.utf8MaxBytes.
> That means that it will fail or will enlarge any ByteBuf that is perfectly
> sized to contain the encoded string.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)