From: "Michael W. Ryder" <_mwryder@...>
Date: 2007-12-06T16:55:34+09:00
Subject: Re: Ruby method to strip out XML codes?

Phrogz wrote:
> On Dec 5, 6:13 pm, "Michael W. Ryder" <_mwry...@worldnet.att.net>
> wrote:
>> I am trying to process an XML file that includes various codes.  The
>> problem I am running into is that some of these codes are inserted into
>> the middle of an encrypted string.  If I display the file using a
>> browser these codes do not show up and copying and pasting the string
>> work fine.  The problem occurs when I try to strip out the string in a
>> program and these "extraneous" XML codes are included.  This of course
>> makes the decryption routine crash.
>> What I am looking for is a simple way to read through the file and
>> remove all the XML codes leaving just plain text.  I could probably
>> write a series of regular expressions to remove each code that I can
>> find in my text but am afraid I might miss some and it will come back to
>> haunt me at a later time.
> 
> str.gsub /</?[^>]+>/, ''
> 
> This will only be a problem if your XML file is legal and has a CDATA
> section which has a literal < character (not &lt;), like:
> 
>    for ( var i=0, len=a.length; i<len; ++i )
> 
> In that case you likely want a proper XML parser (like REXML) and to
> use it.
> 
> Do you really want to remove the XML, or would it suffice to just:
> 
>   str.gsub! '&', '&amp;'
>   str.gsub! '<', '&lt;'
>   str.gsub! '>', '&gt;'
> (and maybe even)
>   str.gsub! '"', '&quot;'
>   str.gsub! "'", '&apos;'
> 
> to make your string valid and escaped for use in an HTML context?

My problem is that the XML file includes &#xD;&#xA; in the middle of a 
couple of fields, especially in the encrypted fields.  If I just strip 
out the encrypted field and try to decrypt it the program crashes as the 
  key is invalid.  I have to remove the "bad" character strings before 
sending it to my decryption program.  I would prefer to do this removal 
before sending the file to my programs so that I don't have to deal with 
these codes.
I assume that the string I am seeing is XML's way of saying CR/LF as DA 
in hex is CR/LF and the output in a browser shows the field being broken 
at that point.  The problem is that is only the ones that I have noticed 
and there may be others hiding in the data.  The XML file is being 
parsed for conversion to our accounts.