[
https://issues.apache.org/jira/browse/AVRO-3748?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17860716#comment-17860716
]
ASF subversion and git services commented on AVRO-3748:
-------------------------------------------------------
Commit 9443fa9b84d4ebf89f0a6dfd7341283609650d98 in avro's branch
refs/heads/main from Oscar Westra van Holthe - Kind
[ https://gitbox.apache.org/repos/asf?p=avro.git;h=9443fa9b8 ]
AVRO-3748: [Java] Fix SeekableInput.skip (#2984)
* AVRO-3748: Fix SeekableInput.skip
Two of the implementations of SeekableInput.skip had a bug: skip was
implemented as seek (i.e. using an absolute input position instead of a
relative one). This fixes that.
* AVRO-3748: Avoid reset+skip confusion
> issue with DataFileSeekableInput.SeekableInputStream.skip
> ---------------------------------------------------------
>
> Key: AVRO-3748
> URL: https://issues.apache.org/jira/browse/AVRO-3748
> Project: Apache Avro
> Issue Type: Bug
> Components: java
> Affects Versions: 1.11.1
> Reporter: Steven Aerts
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.12.0
>
> Time Spent: 1h 10m
> Remaining Estimate: 0h
>
> We found a longstanding bug in the implementation of
> {{DataFileSeekableInput.SeekableInputStream.skip.}}
> This skip function is not hit that often. It can for example be hit when the
> FastReader is enabled and it tries to skip a significant amount of data.
> The implmentation of this function is however fault and can result in data
> corruption or
> {{{}java.io.EOFException{}}}, as instead of skipping the number of bytes, it
> will seek to a wrong place in the file.
>
> We have a pull request ready to fix and test this issue.
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
