W3C home > Mailing lists > Public > xsl-editors@w3.org > July to September 2001

"Bug" in XML Path Language (XPath) Version 1.0 W3C Recommendation 16 November 1999

From: Janko Stamenovic <janko@teletrader.com>
Date: Mon, 9 Jul 2001 06:39:40 -0400 (EDT)
To: <www-xpath-comments@w3.org>, <jjc@jclark.com>, <Steven_DeRose@Brown.edu>
Message-ID: <OLEEKCKMIJGJOKEJDOJCIEBNCBAA.janko@teletrader.com>

Hello,

It seems to me that the specification is a little incomplete in the grammar
part?

I know that it states:

"XPath is intended primarily as a component that can be used by other
specifications. Therefore, XPath relies on specifications that use XPath
(such as [XPointer] and [XSLT]) to specify criteria for conformance of
implementations of XPath and does not define any conformance criteria for
independent implementations of XPath."

But on the other hand, it made confusion to implementors because of:

"[30]    Number    ::=    Digits ('.' Digits?)?
   | '.' Digits
[31]    Digits    ::=    [0-9]+ "

I expect that good implementation MUST behave according to IEEE 754
standard, as in other XML specifications. That this was a real idea of XPath
I conclude from the same document:

"3.5 Numbers
A number represents a floating-point number. A number can have any
double-precision 64-bit format IEEE 754 value [IEEE 754]. These include a
special "Not-a-Number" (NaN) value, positive and negative infinity, and
positive and negative zero. See Section 4.2.3 of [JLS] for a summary of the
key rules of the IEEE 754 standard."

As far as I understand, not handling IEEE 754 numbers IS A BUG in the
implementation, and, sadly, the specification is clearly incomplete (because
of the grammar).

It seems that the grammar shown in XPath spec would not work even for the
simplest cases, since it does not cover negative numbers. Even in errata
 http://www.w3.org/1999/11/REC-xpath-19991116-errata/)  seems that nobody
saw the obvious problem that only adding sign still does not correct the
buggy grammar?

I believe that a correctly written grammar should specify "numbers" as all
possible IEEE formatted numbers, with

a) optional sign, mantissa in which can optionally appear a '.' and optional
positive and negative exponent, separated by 'E' or 'e' from mantisa, or
with
b) (optional sign) 'INF' or with
c) (optional sign) 'NaN'.

Or, as another possibility, it should simply not specify number format at
all and clearly direct to more relevant specification.

I await your opinion in order to be able to

1) Complain to implementors of XPath who do not handle IEEE numbers,

2) Prove in my company that we should stick to the IEEE norms.


Thank you,
Janko Stamenovic
Teletrader AG
Vienna, Austria
Received on Tuesday, 10 July 2001 06:34:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:52 GMT