[
https://issues.apache.org/jira/browse/NIFI-5711?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16659595#comment-16659595
]
ASF GitHub Bot commented on NIFI-5711:
--------------------------------------
Github user markap14 commented on the issue:
https://github.com/apache/nifi/pull/3085
@patricker you do have a good point... interestingly, though, that's not a
bug introduced in your PR, as the call to `readLine()` that existed before
would also never check the max buffer size. the docs say "Specifies the maximum
amount of data to buffer (per file or per line, depending on the Evaluation
Mode)" but unless I'm missing something, it has never been enforced per-line.
Also, I did put up a PR, https://github.com/apache/nifi/pull/3100, that
addresses the performance concern and in that PR, it does actually enforce the
max buffer size (as a 'weak limit'), which until now I had not noticed that we
were previously not doing.
> NLKBufferedReader appears extend and copy portions of the JDK BufferedReader
> ----------------------------------------------------------------------------
>
> Key: NIFI-5711
> URL: https://issues.apache.org/jira/browse/NIFI-5711
> Project: Apache NiFi
> Issue Type: Bug
> Components: Extensions
> Reporter: Joseph Witt
> Assignee: Mark Payne
> Priority: Blocker
> Fix For: 1.8.0
>
>
> NLKBufferedReader appears to extend from then copy then alter JDK code.
> Extension and alter is fine. Copy and paste to get there is not. The JDK
> code is not licensed in a manner conducive for ALv2
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
