David --

I guess the question that the erratum raises is what does

  <xsl:decimal-format />

with no arguments mean.  I just looked at the Recommendation again and I
see what you're saying because there is a default value specified for
each attribute.  I assumed that the default value was only the default
value when the locale was the US or a compatible place but I see what
you're saying that it should always be the default.

Then, what does the Recommendation mean by "The format pattern string is
in a localized notation; the decimal-format determines what characters
have a special meaning in the pattern ..."

I had, mistakenly I think, thought that the localized notation had
something to do with a java Locale.  However, thanks to your repeated,
persistent, and patient reference to the erratum, I now think it means
localized to the stylesheet (and included and imported ones).  This must
be what the part after the semicolon is trying to tell me.

I'll make the corrections in the Java code and clarify to Stephane what
is meant.

Thanks,
Gary


[EMAIL PROTECTED] wrote:
> 
> >> is Stephane Bailliez
> > is Gary Peskin
> 
> >>However what is supposed to be the behavior when a decimal-format does
> >>not define all characters used in the pattern ? In this case it could
> >>simply mean that we decided to go with the locale separators...
> 
> >Yes.  I think that XalanJ should internally keep a list of
> >DecimalFormatSymbols for each named decimal-format and one for the
> >default decimal-format.  This list is initialized by the
> >DefaultFormatSymbols() constructor which constructs a locale-specific
> >list.  This can be overriden by the xsl:decimal-format element.
> 
> This is part of what Erratum E10 clarifies, as I read it.
> http://www.w3.org/1999/11/REC-xslt-19991116-errata#E10
> 
> I think the erratum is trying to say that you never inherit settings from
> the locale; that decimal-separator is '.' worldwide if not specified.
> Therefore the list is initialized by the 10 default values given in
> Section 12.3 of the XSLT spec.
> .................David Marston

Reply via email to