Re: [xsl] [OT] Re: [xsl] Visual XHTML diff

Subject: Re: [xsl] [OT] Re: [xsl] Visual XHTML diff
From: Aaron Gray <aaronngray.lists@xxxxxxxxxxxxxx>
Date: Tue, 29 Dec 2009 16:58:21 +0000
2009/12/29 Xmlizer <xmlizer+xsllist@xxxxxxxxx>
>
> On Mon, Dec 28, 2009 at 6:28 PM, Aaron Gray
> <aaronngray.lists@xxxxxxxxxxxxxx> wrote:
> > 2009/12/28 Aaron Gray <aaronngray.lists@xxxxxxxxxxxxxx>:
> >> 2009/12/28 Michael M|ller-Hillebrand <mmh@xxxxxxxxxxxxx>
> >>>
> >>> Am 24.12.2009 um 23:40 schrieb Aaron Gray:
> >>>
> >>> >> You have both missed my point. I want an XSLT stylesheet to compared
> >>> >> or merge XML(XBEL)/XHTML files resulting in XHTML or XBEL output
> >>> >> files. Maybe I shouldhave not said "Visual".
> >>> >>>>> I am wondering how to do a visual diff of two XHTML files, one
newer
> >>> >>>>> one and a second older one, showing new text in green and
> >>> >>>> old text in
> >>> >>>>> red strike through.
> >>> >>>>> I need to do this in XSLT 1.0 if it is at all possible, although
I
> >>> >>>>> would have thought it was as its a Turin computable problem.
> >>> >>>>>
> >>> >>>>> Many thanks in advance,
> >>> >
> >>> > Anyone give me any clue how to approach this problem ?
> >>>
> >>
> >> Michael,
> >>
> >> Thanks for the reply.
> >>
> >>> Apparently this is not a task that would be very easy to do in XSL,
depending on the level of detail you want in the result. >It would be no fun
at all to implement a longest common string algorithm in XSL 1.0. So, it seems
most solutions use >other tools to prepare something which could then be
turned into a visual representation in a way you like.
> >>
> >> Right, I thought it would be a tough problem if possible at all.
> >>
> >>> I am currently evaluating those options to handle the diff task:
> >>>
> >>> * http://nwalsh.com/java/diffmk/ [not under development]
> >>> * http://xmlunit.sourceforge.net/ [active development]
> >
> > These are Java and I am not running Java on our servers so is not a
> > solution for me :(
> >
> >>> * http://www.xmldifftool.com/ [author says a new version is under
development]
> >
> > This is closed source.
> >
> >> Okay, please let me know what you find. I will have a look too, when I
> >> get the time and right head on.
> >>
> >> I had been considering a DOM based solution too.
> >
> > I think a DOM solution in PHP or Javascript is what I am going to have
> > to consider.
>
> Javascript solution would be better if you want to open source it
> afterward, in case it ends up to be general diff


I am looking at a possible DOM + Hashing + LCS solution. What I am
going to code it in I have not descided yet, probably PHP.
Aaron

Current Thread