5054 2007-09-17 14:13:58 +0000 Unicode character in K2-StringLT-1 2007-09-18 17:29:34 +0000 1 1 1 Unclassified XML Query Test Suite XML Query Test Suite unspecified PC Windows XP CLOSED INVALID P2 normal --- 1 andrew.eisenberg frans.englich public-qt-comments oldest_to_newest 16610 0 andrew.eisenberg 2007-09-17 14:13:58 +0000 Test case K2-StringLT-1 contains the comparison of two large codepoints. I generate the following XQueryX for this test case: <?xml version="1.0"?> <xqx:module xmlns:xqx="http://www.w3.org/2005/XQueryX" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.w3.org/2005/XQueryX http://www.w3.org/2005/XQueryX/xqueryx.xsd"> <xqx:mainModule> <xqx:queryBody> <xqx:ltOp> <xqx:firstOperand> <xqx:stringConstantExpr> <xqx:value>&#60000;</xqx:value> </xqx:stringConstantExpr> </xqx:firstOperand> <xqx:secondOperand> <xqx:stringConstantExpr> <xqx:value>&#55300;</xqx:value> </xqx:stringConstantExpr> </xqx:secondOperand> </xqx:ltOp> </xqx:queryBody> </xqx:mainModule> </xqx:module> When I attempt to validate this XQueryX, I see this error: Character reference "&#55300" is an invalid XML character. I'm weak on the details of Unicode. I believe that character &#55300 is &#xD804. I see the following in http://www.unicode.org/Public/UNIDATA/UnicodeData.txt: D800;<Non Private Use High Surrogate, First>;Cs;0;L;;;;;N;;;;; DB7F;<Non Private Use High Surrogate, Last>;Cs;0;L;;;;;N;;;;; Perhaps you could change &#xD804 to some other character. I've experimented a bit, and &#xD700; validates just fine. 16667 1 mike 2007-09-18 12:57:54 +0000 I think the translation of the query into XQueryX was done incorrectly. From looking at the file at the octet level, the first operand is the octet sequence ee a9 a0, the second is f0 91 85 b0. These are the UTF-8 representations of the characters with codepoints (decimal) 60000 and 70000 respectively. Codepoint 70000 will be represented in UTF-16 as a surrogate pair, and it looks as if your translation has taken the first 16 bits of the surrogate pair as representing the entire character. 16671 2 andrew.eisenberg 2007-09-18 17:29:27 +0000 Mike, your comment helped me pinpoint the bug in the XQueryX generation. I agree that the test case is correct as it is.