[Schematron] Help sought: implementation of Character Repertoire in XSLT2 for embedding in schematron

Lars Huttar lars_huttar at sil.org
Tue Sep 23 00:05:59 EDT 2008


On 9/20/2008 9:15 AM, Dave Pawson wrote:
> 2008/9/20 Rick Jelliffe <rjelliffe at allette.com.au>:
>
>   
>> There needs to be a leading and trailing ^ and $ to get the match for the
>> whole string.   But it would be better to go through character-by-character
>> to identify the particular unexpected character.
>>     
>
>
> That bothers me; processing a single character at a time is grossly inefficient.
>
> Ideally process a document in XML,
> next best thing a string.
>
> Otherwise the processing time is going to send us to sleep!
>
>
> regards
>
>   
I confess I haven't been following this too closely; but couldn't you do 
both, i.e. use "^...$" to determine efficiently whether the pattern 
matches or not, and then if it doesn't (hopefully a rare case), go 
through character-by-character to find the offending one?


Lars




More information about the Schematron mailing list