RE: saxon fatal error reports odd location

Firstly, credit where credit is due: the XML parser that ships with Saxon is
David Megginson's AElfred, upgraded to SAX2 by David Brownell, and with two
or three bug fixes added by me. 

In most cases where it gets the line numbering wrong, it's because it's
confused about the character encoding. It's difficult to count newline
characters if you're looking for them in the wrong encoding.

That doesn't seem to be the case here. I think it's more likely to be due to
differences in handling CR, CRLF, and LF sequences. Looking at the code
(which I'm
hesitant to touch as I don't have a regression test suite for it) I suspect
that it's not incrementing the line number and resetting the column number
when it finds a CR that isn't paired with an LF, while processing PCDATA.

There's some pretty hairy optimisation code in AElfred doing lookahead and
backtracking, and I suspect speed was regarded as more important than
getting the column number right.

Mike Kay
> 
> Fatal error reported by XML parser: required string (expected 
> "BIOGPROSE")
>   URL:    file:/C:/Documents and Settings/dave/Desktop/XSLT/all.xml
>   Line:   27643
>   Column: 1982
> Transform failed:  required string (expected "BIOGPROSE")
> 
> I considered posting this message at 
http://saxon.xsl.listbot.com/, but I
couldn't find a way to post to it without actually registering and
subscribing, which I presumed would flood my mailbox 

Traffic on that list is quite bearable, about 1 to 5 messages on a typical
day.

Mike Kay


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list