iwextattr -x and utf-8 keys

jed

I have a metadata key/value with utf-8 data that I am setting using the -x option. Everything is fine and dandy when I get the output using iwextattr <vpath>, but the key is not correctly encoded when retrieved with the "-x" option. Any ideas before I open a support case?

Example:
VALID:
/app/suncom/iw-home/bin/iwextattr /default/main/www.sun.com/web_site/en/content/WORKAREA/draft/books/md.foo
&#26085;&#26412; (Japan)=&#26085;&#26412; (Japan)

VALID:
/app/suncom/iw-home/bin/iwextattr -e /default/main/www.sun.com/web_site/en/content/WORKAREA/draft/books/md.foo
&#26085;&#26412; (Japan)

INVALID:
/app/suncom/iw-home/bin/iwextattr -x /default/main/www.sun.com/web_site/en/content/WORKAREA/draft/books/md.foo
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE ealist
[
<!ELEMENT ealist (ea)*>
<!ELEMENT ea (key, value*)>
<!ELEMENT key (#PCDATA)>
<!ELEMENT value (#PCDATA)>
<!ATTLIST value encoding (text|hex) "text">
]
>

<ealist>
<ea>
Warning invalid XML character 0xffffffe6 in ea key &#26085;&#26412; (Japan)
Character not encoded into key.
Warning invalid XML character 0xffffff97 in ea key &#26085;&#26412; (Japan)
Character not encoded into key.
Warning invalid XML character 0xffffffa5 in ea key &#26085;&#26412; (Japan)
Character not encoded into key.
Warning invalid XML character 0xffffffe6 in ea key &#26085;&#26412; (Japan)
Character not encoded into key.
Warning invalid XML character 0xffffff9c in ea key &#26085;&#26412; (Japan)
Character not encoded into key.
Warning invalid XML character 0xffffffac in ea key &#26085;&#26412; (Japan)
Character not encoded into key.
<key>&#x20;(Japan)</key>
<value>&#26085;&#26412;&#x20;(Japan)</value>
</ea>
</ealist>


I set the metadat using the following:
/app/suncom/iw-home/bin/iwextattr -s -x "/default/main/www.sun.com/web_site/en/content/WORKAREA/draft/books/md.foo" < "xml.data"

xml.data is the following:
<eaList><ea><key>&#26085;&#26412; (Japan)</key><value>&#26085;&#26412; (Japan)</value></ea></eaList>


--
Jed Michnowicz
jedm@sun.com
Content Management Engineering
Sun Microsystems

Bill Klish

We had this same problem. You need to use the -f switch for this to work with UTF-8. We have done this before in past implementations (5.5 and 6.1).

Just flipping through the 6.5 CLT manual, it looks like there is a -u switch now which says the following:
UTF8 encode extended attribute values on input, and decode on output.

I know -f works because we use it, not sure about the -u, but it looks promising.

After re-reading, you are asking about retriving UTF-8 encoded keys and not values. Hmm, not sure about that.

Edited by bk99nd on 01/05/06 02:25 PM (server time).