Hi Thomas,
You reminded me that I forgot to answer your question about
work-arounds. I'm sorry to say that I don't have one. Does anyone out
there have any suggestions?
Thanks,
Henry
------------------------------------------------------------------
Henry Zongaro Xalan development
IBM SWS Toronto Lab Tie Line 969-6044; Phone (905) 413-6044
mailto:[EMAIL PROTECTED]
Henry Zongaro/Toronto/[EMAIL PROTECTED]
11/28/2002 06:04 AM
To: "Thomas Dandanell Berg" <[EMAIL PROTECTED]>, [EMAIL PROTECTED]
cc:
Subject: Re: Vedr: Re: Reg.: Xalan-Java version 2.4.1
Hi Thomas,
I hope you don't mind that I'm responding on the mailing list, so
that others can see this information.
Sorry for not being more explicit: yes, you're right, the Collator
is the Java object that is used to perform culturally-specific sorting. I
took a look at the Java bug database, and found the following bug:
http://developer.java.sun.com/developer/bugParade/bugs/4174436.html. It
mentions precisely your problem with the way v and w are being treated for
Danish.
I will append a comment to that bug, which has been open for several
years, to let them know that users are still being affected by this.
Thanks,
Henry
------------------------------------------------------------------
Henry Zongaro Xalan development
IBM SWS Toronto Lab Tie Line 969-6044; Phone (905) 413-6044
mailto:[EMAIL PROTECTED]
"Thomas Dandanell Berg" <[EMAIL PROTECTED]>
11/28/2002 04:15 AM
To: Henry Zongaro/Toronto/[EMAIL PROTECTED]
cc:
Subject: Vedr: Re: Reg.: Xalan-Java version 2.4.1
I get the exact result as u did, and that is the problem in my opinion. As
i understand you the Java Collator(do not know what it is!) is the
algorithm which specifies the sort-order, did i get that right?
It might be right that the parser parses according to the algorithm, but i
think the algorithm is wrong:
your statement:
"the difference between 'a' and 'b' is considered to be more significant
than the difference between 'W'
and 'V'."
In Denmark we sort no different than in England according to "v" and "w"!
This illustrate the difference (in my opinion the bug):
xsl:sort lang="da" xsl:sort lang="en"
ba ba
Ba Ba
bb bb
Bb Bb
ca ca
Ca Ca
cb cb
Cb Cb
va va
Va Va
wa vb
Wa Vb
vb wa
Vb Wa
wb wb
Wb Wb
As you can see there is a difference in the sorting order, according to
v/V and w/W words,
between the Danish(da) and the English(en) sort!
The parser should sort the above words exactly in the same order no matter
if lang="da" or lang="en".
Therefore i believe there is a bug en the algorithm!
Do you see my point?
Is the concrete sort routine implemented in Xalan, standard Java or
somewhere else?
Is there a work-around? maybe a way to explicitly set the sort order in
some sort of inputfile?
Thanks a lot for your help so far ;-�
regards
Thomas
