- From: <scott_boag@us.ibm.com>
- Date: Sun, 26 Jan 2003 22:44:43 -0500
- To: "Twan van Laarhoven" <twanvl@hotmail.com>
- Cc: public-qt-comments@w3.org
Hi Twan. Thank you for the report. public-qt-comments-request@w3.org wrote on 01/23/2003 11:19:48 AM: > if (A) then B instance of element C context type else/D > According to precedence rules, should be interpreted the same as: > if (A) then (B instance of element C context type else/D) > However, this is not a valid expression, since there is no 'else' in if > expression. Right, and neither is the first one. The lex rules (admitting that there were some massive bugs in the lex tables in the last public draft) determine that "else" occurs in a position where a QName is expected, as opposed to the keyword "else" expected in the OPERATOR state. The following parses fine, and as expected: if (A) then B instance of element C context type else/D else /D So, at least in the current draft (as opposed to the last public draft), I don't see any problem. Also double-checked that your: > if (A) then (B instance of element C context type) else (/D) Parses OK. It does. Note there are some ambiguity bugs with Sequence type for keywords vs. QNames, such as "element" etc. We're working on a solution for this. Thank you again for your feedback! -scott
Received on Sunday, 26 January 2003 22:43:52 UTC