[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]
Re: [xml] XPath mod operator problems
- From: Daniel Veillard <veillard redhat com>
- To: Geert Kloosterman <geertk ai rug nl>
- Cc: xml gnome org
- Subject: Re: [xml] XPath mod operator problems
- Date: Thu, 27 Jun 2002 12:40:51 -0400
On Thu, Jun 27, 2002 at 05:57:21PM +0200, Geert Kloosterman wrote:
> Hi Daniel, all,
>
> Part of this has allready been posted to the xslt list, but since the
> real problem seems to be within libxml2, I'll continue here.
yup !
> I can't reach the bugzilla site at the moment, so maybe this
> phenomenon has allready been posted.
>
> I've used libxml2-2.4.22 on all systems mentioned.
>
> Here we go. On i386 platform I get the correct results:
>
>
> ./testXPath --expr '8 mod 3 = 2'
> Object is a Boolean : true
>
>
> But on HP-UX 10.20 and linux-alpha:
>
>
> ./testXPath --expr '8 mod 3'
> Object is a number : 2
>
> ./testXPath --expr '8 mod 3 = 2'
> Object is a Boolean : false
>
>
> Can anybody reproduce this on other platforms?
> Daniel, should this test be added to the regression tests?
>
yes, I take patches :-)
> When I change the following lines of xpath.c (l. 4890, 4891),
> xmlXPathModValues:
>
> tmp=arg1/arg2;
> ctxt->value->floatval = arg2 * (tmp - (double)((int)tmp));
>
> into
>
> ctxt->value->floatval = fmod(arg1, arg2);
>
> the behaviour looks correct. But I don't know if fmod conforms
> to the XPath requirements of the `mod' operator.
And I also don't know the portability of fmod (and being on the
road I don't have my books to check :-/ )
Maybe someone with access to multiple platforms could check this
Bjorn for example ?
Daniel
--
Daniel Veillard | Red Hat Network https://rhn.redhat.com/
veillard redhat com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/
[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]