[Phpslash-devel] Re: [Back-end-development] Annoying Problem with stripBadHTML

[Mike G. <mi...@op...>]

Hello Sam,

I've cc'd the slash development list.

On Thu, 2003-06-19 at 23:14, Sam Williams wrote:
> On Fri, 2003-06-20 at 02:38, Mike Gifford wrote:
> > In cutting/pasting it in sections it seemed to get hung up on:
> > 	<a 
> > 	name=Baird></a><br>
> > After removing all of these hanging <a's I was able to paste it in
> > without difficult.  I just had to remove a number of them.
> I seem to remember having problems with this function in phpSlash a
> while back.
> At a guess, I would say that the newline, LF or LF/CR character pair
> (which it is depends on whether the text came from Windows, Unix or Mac)
> after the <a are messing up the regex matching. 

Yup.  

> Looking at (an old version of) the function, I don't think the lines
> designed to 'standardise' html tags will match a hanging tag:
> 
> $str = eregi_replace("<[[:space:]]*([^>]*)[[:space:]]*>","<\\1>",$str);
> $str = eregi_replace("<a([^>]*)href=\"?([^\"]*)\"?([^>]*)>",
> 	"<a href=\\2>", $str);

This is from the phpSlash CVS version:
$str = eregi_replace("<[[:space:]]*([^>]*)[[:space:]]*>","<\\1>",$str);
$str = eregi_replace("<a([^>]*)href=\"?([^\"]*)\"?([^>]*)>",
                         "<a href=\\2>", $str);

> and I don't think this will either:
> while (eregi("<([^> ]*)([^>]*)>",$str,$reg)) {

I'm not a regex expert, but I can't see why this would be interrupted by
a line break...  

> A quick fix would be another eregi_replace that strips all newline, LF
> and CR characters from within tags before anything else is done with the
> string.

Yes, but that would make it very difficult to edit afterwards, right?

> Perhaps it's time to update the stripBadHTML so that it uses the native
> PHP function that was introduced in 3.0.8: 

I'd be in favour of moving this way.  I would think it would speed up
the code to have the processing done by native php functions.

> string strip_tags ( string str [, string allowable_tags])   

What it doesn't allow is the degree of control that stripBadHTML
presently offers.  The ability to allow tags or tags and definitions is
nice.  Mind you I don't know how many folks use that.

> I'm not sure how well this works compared to the phpslash code, but it
> is much more readable :-). This is probably something for the
> phpSlash-devel list...

It doesn't seem to be a problem in phpSlash.  the line break doesn't
seem to interfere with the code.

Mike
-- 
Mike Gifford, OpenConcept Consulting
Free Software for Social Change -> http://www.openconcept.ca
Featured Client: CUPE National -> http://www.cupe.ca
Whoever controls the media-the images-controls the culture - A. Ginsberg