Distribution/Version: Gentoo Linux x86

Parsing a document using the SAX API, an "&" entity in the value of an 
attribute is returned to the start element handler as "&".  This is a 
valid character entity for an ampersand, but I was expecting the parser to 
return "&" at this point.  I've confirmed that "<" and ">" in an 
attribute value are returned as "<" and ">" by the start element handler.

Sample input document:

<a>
        <b c="Tim &amp; Bev">Tim &amp; Bev</b>
        <b c="1 &lt; 2">1 &lt; 2</b>
        <b c="4 &gt; 3">4 &gt; 3</b>
        <b c="Don&apos;t!">Don&apos;t!</b>
        <b c="Call me &quot;Ishmael&quot;">Call me &quot;Ishmael&quot;</b>
</a>

Here is the data I'm getting from the parser (output using my own code for 
debugging purposes, this is obviously *not* valid XML):

<a>
        <b c="Tim &#38; Bev">Tim & Bev</b>
        <b c="1 < 2">1 < 2</b>
        <b c="4 > 3">4 > 3</b>
        <b c="Don't!">Don't!</b>
        <b c="Call me "Ishmael"">Call me "Ishmael"</b>
</a>

Note that this only affects &amp; and only if it is an attribute value - an 
ampersand in element text is returned to the caller as "&" as expected.

Tim Shead
tshead@k-3d.com