From: Ivor Paul <ivorpaul@...>
Date: 2007-07-20T03:27:49+09:00
Subject: Re: hpricot question

------=_Part_115695_4675520.1184869670889
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

Thanks peter - that is what I wanted to hear! I knew there was some reason
like that.

On 7/19/07, Peter Cooper <peter@peterc.org> wrote:
>
> On 7/19/07, Ivor Paul <ivorpaul@gmail.com> wrote:
> > I'm a total hpricot newbie, so this is only out of curiousity...
> >
> > Is the following expected behaviour?
> > I found this while struggling through a nasty xml parsing bug.
> >
> > hstr = Hpricot('<a/><b>c</b>').to_s #or to_html
> >
> > returns => '<a><b>c</b></a>'
> >
> > I have found that using to_original_html returns '<a/><b>c</b>' as
> expected
> > but I was wondering why you would want <a/> to wrap the other contents
> when
> > it obviously isn't correct? Is this handy behaviour for scrapping?
>
> Yeah, I never understand that either. But it's easily fixed:
>
> h = Hpricot('<a/><b>c</b>', :xml => true)
> h.to_s =>  "<a></a><b>c</b>"
>
> I think it's because when you don't define :xml as true, then Hpricot
> acts as an *HTML* processor, and therefore tries to correct bad HTML..
> and somehow ends up with the badness of wrapping the empty element.
>
> Cheers,
> Peter Cooper
> http://www.rubyinside.com/
>
>

------=_Part_115695_4675520.1184869670889--