Re: [xslt] format-number inconsistent with java results

[Ken Neighbors <ken nsds com>]

Bjorn,

Thanks for your detailed feedback on the interpretation of the JDK.  I
agree that these special cases seem to be either illegal or fall
outside of the spec, and also wish XSLT had spelled it out better
instead of relying on JDK.  I interpret your last comment as saying we
should go ahead anyway and keep the patch as an extension to the spec.

I just took a look at the XSLT 2.0 draft, and it specifies
format-number in much more detail without referring to JDK, so it
seems they are rectifying their mistake (as you put it).  In
particular, it allows '.0' and '1.'.  A little surprisingly, I think
it also allows '.'!  Here is the relevant section:

  http://www.w3.org/TR/xslt20/#section-Processing-the-Picture-String

Ken