[
https://issues.apache.org/jira/browse/IO-400?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Henri Yandell moved LANG-897 to IO-400:
---------------------------------------
Affects Version/s: (was: 3.1)
Key: IO-400 (was: LANG-897)
Project: Commons IO (was: Commons Lang)
> IOUtils: add support for copying from large byte buffers
> --------------------------------------------------------
>
> Key: IO-400
> URL: https://issues.apache.org/jira/browse/IO-400
> Project: Commons IO
> Issue Type: New Feature
> Reporter: Sebb
>
> Trying to write a large byte array to a FileOutputStream may cause OOME.
> This is because such output requires the use of native code, and native code
> may need to copy the array in order to access it safely, see:
> http://docs.oracle.com/javase/7/docs/technotes/guides/jni/spec/design.html#wp1265
> It might therefore be useful to have a method which writes from the byte
> array in chunks. One can create a ByteArrayInputStream from the input, and
> then use one of the copy() methods, but that creates an unnecessary array
> buffer (albeit only 4k).
> There are already write methods which copy byte[] to OutputStream and char[]
> to Writer.
> Some or all of these could be converted to use chunked output, or there could
> be new methods to implement the chunking.
> Here is a sample implementation of a stand-alone method:
> {code}
> public static void writeChunked(byte[] data, OutputStream output) throws
> IOException {
> int bytes = data.length;
> int offset = 0;
> while(bytes > 0) {
> int chunk = Math.min(bytes, DEFAULT_BUFFER_SIZE);
> output.write(data, offset, chunk);
> bytes -= chunk;
> offset += chunk;
> }
> }
> {code}
--
This message was sent by Atlassian JIRA
(v6.1#6144)
