Benson, perhaps "contrived" would have been better than "hypothetical"
:). That being said, I also hadn't thought about other JVM implementations.
I wonder if leaving a commented note in the accumulo-env.sh script for
alternative namings for the "file.encoding" name and the JVM it applies
to would be sufficient?
David, can you give some sort of feel for the usages of the getBytes()
calls? Since most of the API deals with things in terms of Text and
byte[] (Key and Value decomposed), are most of the usages
configuration/user-input based as your initial snippet from
InputFormatBase showed?
On 10/29/2012 12:42 PM, John Vines wrote:
Are there any experts when it comes to character encodings? First of all, I
would like to make sure there are no sacrifices being made by forcing UTF-8.
From there, if I think JVM properties is the way to go. Should there be ANY
sort of shortfall with UTF-8, we should allow users to switch the encoding
to the type of their pleasure. We can tweak the scripts to set the jvm
property but still allow users to override should they need it in their
setup. This allows us to not only avoid a massive code change, it also
makes it easier for users to switch to an encoding should they have a need
to.
John
On Mon, Oct 29, 2012 at 12:24 PM, Benson Margulies <[email protected]>wrote:
On Mon, Oct 29, 2012 at 12:21 PM, Josh Elser <[email protected]> wrote:
David, I beg to differ.
Setting it via the JVM property is a single change to make, whereas if
you
change every single usage of getBytes(), you now forced the next person
to
branch the code, change everything to UTF16 (hypothetical use case) and
continue a diverged codebase forever.
Typically, the reason(s) that people don't take this approach are:
a: a fear that other JVMs don't have this parameter, or don't have it
under the same name.
b: a desire to read or write files for uses in 'the platform encoding'
whatever it is, in addition to whatever needs to be done in UTF-8.
I'd be very surprised if Accumulo ever decided to do this sort of
thing in UTF-16.
I would say that the reason that such a JVM property exists is to
alleviate
you from having to make these code changes in the first place.
On 10/29/2012 12:00 PM, David Medinets wrote:
I like the idea of making the change explicit in the source code.
Setting the encoding in the jvm property would be easier but not as
explicit. I have a few dozen of the files changed. Today I have free
time since Hurricane Sandy has closed offices.
On Mon, Oct 29, 2012 at 11:39 AM, William Slacum
<[email protected]> wrote:
Isn't it easier to just set the JVM property `file.encoding`?
On Sun, Oct 28, 2012 at 3:18 PM, Ed Kohlwey <[email protected]>
wrote:
If you use a private static field in each class for the charset, it
will
basically be a singleton because charsets are cached in char
set.forname.
IMHO this is a somewhat cleaner approach than having lots of static
imports
to utility classes with lots of constants in them.
On Oct 28, 2012 5:50 PM, "David Medinets" <[email protected]>
wrote:
https://issues.apache.org/jira/browse/ACCUMULO-241?focusedCommentId=13449680&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13449680
In this comment, John mentioned that all getBytes() method calls
should be changed to use UTF8. There are about 1,800 getBytes() calls
and not all of them involve String objects. I am working on ways to
identify a subset of these calls to change.
I have created https://issues.apache.org/jira/browse/ACCUMULO-836 to
track this issue.
Should we create one static Charset object?
Class AccumuloDefaultCharset {
public static Charset UTF8 = Charset.forName("UTF8");
}
Should we use a static constant?
public static String UTF8 = "UTF8";
I have found one instance of getBytes() in InputFormatBase:
protected static byte[] getPassword(Configuration conf) {
return Base64.decodeBase64(conf.get(PASSWORD, "").getBytes());
}
Are there any reasons why I can't start specifying the charset? Is
UTF8 the right Charset to use? I am not an expert in non-English
charsets, so guidance would be welcome.
