On 27/02/2013 15:56, John English wrote:
On 27/02/2013 12:41, Michael Kay wrote:
Note that the reason some parsers fail to provide location
information for an
encoding error is that in a properly layered architecture you need to do
decoding before you start recognizing line endings, which means that the
information about line numbers is typically not available at the
point where the
decoding error is detected.
I'm not sure I understand this. It seems from what you're saying that,
for example, a compiler should also be unable to give line number
information for errors (unless it's not "properly layered"). Shouldn't
the tokeniser keep track of where each token comes from in the source
file and be able to tell you this if asked?
My point is that to break a file into lines you need first to decode the
bytes, and if it's sufficiently corrupt that you can't decode it, then
knowing what line you're on requires some fancy footwork. Your decoding
error was happening well before tokenization.
Michael Kay
Saxonica
--~------------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
To unsubscribe, go to: http://lists.mulberrytech.com/xsl-list/
or e-mail: <mailto:xsl-list-unsubscribe(_at_)lists(_dot_)mulberrytech(_dot_)com>
--~--