This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Certain tests have inconsistent XQueryX versions. For example the XQuery version of K-RangeExpr-31 is: (remove((2.e0, 4), 1) treat as xs:integer to 4) Whereas the equivalent XQueryX query says (more or less): (remove((2.e0, 4), 1) treat as xs:integer to 4) eq 4 which obviously gives a different result. This is the list of tests I have noticed similar problems with so far, and the problems with the XQX queries: K-RangeExpr-31 K-RangeExpr-32 K-WhereExpr-1 K-WhereExpr-2 K-FunctionProlog-19 K-FunctionProlog-21 K-FunctionProlog-22 K-FunctionProlog-23 K-FunctionProlog-39 K-FunctionProlog-40 K-FunctionProlog-48 - see bug 4778 K-FunctionProlog-49 - missing empty() K-FunctionProlog-50 - missing argument to myFunction K-FunctionProlog-51 - function body is error(), not $arg K-FunctionProlog-52 - XQX version is that of K-FunctionProlog-53 K-FunctionProlog-53 - XQX version is that of K-FunctionProlog-54 K-FunctionProlog-55 - XQX version is that of K-FunctionProlog-56 K-FunctionProlog-56 - XQX version is that of K-FunctionProlog-57 K-FunctionProlog-58 - XQX version is that of K-FunctionProlog-59 K-FunctionProlog-59 - XQX version is that of K-FunctionProlog-60 K-FunctionProlog-60 - XQX version is that of K-FunctionProlog-61 K-FunctionProlog-61 - XQX version is that of K-FunctionProlog-62 K-FunctionProlog-62 - XQX version is that of K-FunctionProlog-63 K-FunctionProlog-63 - XQX version is that of K-FunctionProlog-64 K-FunctionProlog-64 - XQX version is that of K-FunctionProlog-65 K-FunctionProlog-65 - XQX version is that of K-FunctionProlog-66 K-FunctionProlog-66 - XQX version is that of K-FunctionProlog-67 K-FunctionProlog-67 - XQX version is that of K-FunctionProlog-68 K-SeqExprCastable-21 K-SeqExprCastable-36 In the first group of tests, there is an extra <eqOp> around the test, as above. The <xqx:singleType> elements in K-SeqExprCastable-21 and K-SeqExprCastable-36 are missing <xqx:optional /> tags to keep it in line with the xquery versions (should be xs:QName? and xs:integer? not xs:QName and xs:integer)
More tests that suffer from similar problems: K-TraceFunc-4 - has an extra <xqx:eqOp> K-CodePointToStringFunc-26 - is actually K-CodePointToStringFunc-29 K-StringToCodePointFunc-8 - extra <xqx:eqOp> K-StringToCodePointFunc-14 - extra <xqx:eqOp> K-StringToCodePointFunc-15 - extra <xqx:eqOp> K-StringToCodePointFunc-16 - extra <xqx:eqOp> K-StringToCodePointFunc-17 - extra <xqx:eqOp> K-StringToCodePointFunc-18 - extra <xqx:eqOp> K-StringToCodePointFunc-19 - extra <xqx:eqOp> K-StringToCodePointFunc-20 - extra <xqx:eqOp> K-StringToCodePointFunc-21 - extra <xqx:eqOp> K-EncodeURIfunc-4 - extra <xqx:eqOp> K-SeqIndexOfFunc-6 - extra <xqx:eqOp> K-SeqIndexOfFunc-12 - extra <xqx:eqOp> K-SeqIndexOfFunc-14 - extra <xqx:eqOp> K-SeqIndexOfFunc-17 - extra <xqx:eqOp> K-SeqReverseFunc-10 - extra <xqx:eqOp> K-SeqReverseFunc-11 - extra <xqx:eqOp> K-SeqReverseFunc-17 - extra <xqx:eqOp> K-SeqReverseFunc-18 - extra <xqx:eqOp> K-ContextPositionFunc-2 - extra <xqx:eqOp> K-ContextLastFunc-2 - extra <xqx:eqOp> K-ContextLastFunc-29 - extra <xqx:eqOp>
I was testing against the CVS version, not version 1.0.2, so most of these bugs are invalid. However the following two cases are inconsistent in version 1.0.2: K-SeqExprCastable-21 K-SeqExprCastable-36 The <xqx:singleType> elements in K-SeqExprCastable-21 and K-SeqExprCastable-36 are missing <xqx:optional /> tags to keep them in line with the xquery versions (should be xs:QName? and xs:integer? not xs:QName and xs:integer)
The following tests are also missing <xqx:optional> tags: K2-SeqExprCast-2 K2-SeqExprCastable-21 K2-SeqExprCastable-36
Our XQueryX tests are generated from our XQuery tests. I've been a bit lax, and this hasn't been done in quite a while. I've just refreshed the XQueryX tests, and I believe that this will address the many problems that you've listed. Apologies for putting you through this. I'll try not to let them get so out of synch in the future. Please close this bug report if you agree with this resolution.