Sean Owen created SPARK-13823:
---------------------------------
Summary: Always specify Charset in String <-> byte[] conversions
(and remaining Coverity items)
Key: SPARK-13823
URL: https://issues.apache.org/jira/browse/SPARK-13823
Project: Spark
Issue Type: Improvement
Components: Spark Core, SQL, Streaming
Affects Versions: 2.0.0
Reporter: Sean Owen
Assignee: Sean Owen
Priority: Minor
Most of the remaining items from the last Coverity scan concern using, for
example, the constructor {{new String(byte[])}} or the method
{{String.getBytes()}}, or similarly for constructors of {{InputStreamReader}}
and {{OutputStreamWriter}}. These use the platform default encoding, which
means their behavior may change in different locales, which should be
undesirable in all cases in Spark.
It makes sense to specify UTF-8 as the default everywhere; where already
specified, it's UTF-8 in 95% of cases. A few tests set US-ASCII, but UTF-8 is a
superset.
We should also consistently use {{StandardCharsets.UTF_8}} rather than "UTF-8"
or Guava's {{Charsets.UTF_8}} to specify this.
(Finally, we should touch up the other few remaining Coverity scan items, which
are trivial, while we're here.)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
