[
https://issues.apache.org/jira/browse/MAPREDUCE-7333?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17313408#comment-17313408
]
Marcono1234 commented on MAPREDUCE-7333:
----------------------------------------
That would probably work as well, though I don't think that a call to
{{Integer.toHexString}} will be a performance problem.
If you decide to use the array approach, it might be good to:
- Make the array a {{static final}} field to avoid creating it every time the
method is called. Though maybe using a local array is actually more performant?
I am not very familiar with performance tuning in this regard; maybe the
compiler or the JVM performs escape analysis and only allocates the array on
the stack.
- Storing only {{char}} inside the array (instead of {{String}}); StringBuilder
has an {{append(char)}} method which can then be used
> SecureShuffleUtils.toHex(byte[]) creates malformed hex string
> -------------------------------------------------------------
>
> Key: MAPREDUCE-7333
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-7333
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Affects Versions: 3.2.2
> Reporter: Marcono1234
> Priority: Major
>
> {{org.apache.hadoop.mapreduce.security.SecureShuffleUtils.toHex(byte[])}}
> creates malformed hex strings:
> {code}
> for (byte b : ba) {
> ps.printf("%x", b);
> }
> {code}
> The pattern {{"%x"}} would for bytes < 16 only have on hex char and for
> example both {{1, 0}} and {{16}} would have the result {{"10"}}.
> A correct (and more efficient) implementation would be:
> {code}
> public static String toHex(byte[] ba) {
> StringBuilder sb = new StringBuilder(ba.length * 2);
> for (byte b : ba) {
> int unsignedB = b & 0xFF;
> if (unsignedB < 16) {
> sb.append('0');
> }
> sb.append(Integer.toHexString(unsignedB));
> }
> return sb.toString();
> }
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: mapreduce-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: mapreduce-issues-h...@hadoop.apache.org
