Unicode matchChunk (was multiple-word phrases, links)

[Kjetil Rå Hauge]

At 13:59 +0200 12-09-03, Toma Tasovac wrote:
>I've looked a little more into this.  The 
>problem is not necessarily with the matchChunk 
>function, but with uniencode.  Namely, if the 
>text is typed directly into the field, the 
>matchChunk function will work.  If, however, I 
>use the uniencode function to put the same text 
>file into the field, the matchChunk will fail. 
>Anyway, it's in bugzilla now.
>
>Am Donnerstag, 11.09.03 um 19:54 Uhr schrieb Tuviah Snyder:
>
>>>Chris, your script works great with English, but I can't get it to work
>>>with Unicode text.
>>Maybe try UTF8ing the text first before sending to the matchChunk function.
>>

It may seem that matchChunk likes utf-16 better. 
I was able to make the function work in both 
fields in your test stack in Bugzilla by copying 
the text from field importedText, pasting into 
TextEdit, making sure it was in the Lucida Grande 
font, saving in TextEdit as utf-16 text, and 
reading in as binary:
  answer file "Please choose a file:"
  if it is not empty then put "binfile:"&it into myPath
  set the unicodeText of field "importedtext" to URL myPath

(And if you fix the typo "vesteckt sein" in the 
other field you will get a full set of matches in 
both fields...)
-- 
--- Kjetil Rå Hauge, U. of Oslo. Tel. +47/22856710, fax +47/22854140
--- (this msg sent from home, +47/67148424, fax +1/5084372444)