commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Niall Pemberton (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (IO-277) ReaderInputStream enters infinite loop when it encounters an unmappable character
Date Mon, 05 Sep 2011 23:40:09 GMT

     [ https://issues.apache.org/jira/browse/IO-277?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Niall Pemberton resolved IO-277.
--------------------------------

       Resolution: Fixed
    Fix Version/s: 2.1
         Assignee: Niall Pemberton

I have implemented your suggestion to replace unmappable characters so that behaviour is consistent
with WriterOutputStream. Also I have added constructors to ReaderInputStream/WriterOutputStream
that take CharsetEncoder/CharsetDecoder respectively so that if this is not the desired behaviour,
then people can define their own. Thanks for reporting this and the test case. 

> ReaderInputStream enters infinite loop when it encounters an unmappable character
> ---------------------------------------------------------------------------------
>
>                 Key: IO-277
>                 URL: https://issues.apache.org/jira/browse/IO-277
>             Project: Commons IO
>          Issue Type: Bug
>          Components: Streams/Writers
>    Affects Versions: 2.0.1
>            Reporter: Mike Thomas
>            Assignee: Niall Pemberton
>             Fix For: 2.1
>
>         Attachments: TestReaderInputStreamLoop.java
>
>
> The ReaderInputStream.read(byte[] b, int off, int len) method enters an infinite loop
when its CharsetEncoder encounters an unmappable character in the input buffer.
> When its CharsetEncoder encounters an unmappable character, the value of CoderResult
lastCoderResult.isUnmappable() == true, and Reader.read() is not invoked on the underlying
Reader ever again.
> Attaching source file that reproduces this behavior.
> One fix to consider is to call CharsetEncoder.onUnmappableCharacter(CodingErrorAction)
in the ReaderInputStream constructor with a value other than the default CodingErrorAction.REPORT.
e.g.:
> public ReaderInputStream(Reader reader, Charset charset, int bufferSize) {
>             this.reader = reader;
>             encoder = charset.newEncoder();
>             encoder.onUnmappableCharacter(CodingErrorAction.REPLACE);
> ...
> By replacing the unmappable character with encoder's default replacement character, this
effectively prevents the infinite loop from occurring. I'm not sure if that's the ideal behavior,
but it seems fairly consistent with what org.apache.commons.io.output.WriterOutputStream does.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message