Lists Home |
Date Index |
On Sun, Feb 16, 2003 at 09:54:32AM -0800, Tim Bray wrote:
> Well, XML1.1 is moving in that direction. Even given that, I think that
> XML 1.0's approach, with a big table right in the spec saying "here are
> the legal characters", was probably correct; I (and I'm sure many other
> programmers) ran a perl script over the spec to extract the char parsing
> tables. -Tim
I used vi regexps directly, and recorded those in the C source file :-) !
:1,$ s/\[#x\([0-9A-Z]*\)-#x\([0-9A-Z]*\)\]/ (((c) >= 0x\1) \&\& ((c) <= 0x\2)) ||/
:1,$ s/#x\([0-9A-Z]*\)/ ((c) == 0x\1) ||/
of course the result was later modified a bit to speed up the test.
In order to try to turn a useless post into an useful one, did someone
tried to implement the character normalization checking of XML-1.1 ?
I looked at the ICU sample code a few months ago and this simply scared
me mostly due to my perception of that code complexity and runtime cost.
Daniel Veillard | Red Hat Network https://rhn.redhat.com/
email@example.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/