Re: [xsl] Preserving numeric character entity reference

[Abel Braaksma <abel.online@xxxxxxxxx>]

Kjetil Kjernsmo wrote:
> I can't use the us-ascii output encoding, as is usually suggested in 
> cases like these, because, the numeric character entity references are 
> actually in the us-ascii range anyway (and we want UTF-8 for the rest). 
>
> Is this doable?
>
> Are there any other obfuscation techniques that can be done easily in 
> XSLT that you can suggest?
>   


Hi Kjetil,

You didn't specify the version of XSLT you are using, but here are a 
couple of ways that work with XSLT 2:

 1. Use xsl:character-map to create literal strings like   for 
space and A for A, without the XML serializer interfering with it.
 2. Use strings like   if obfuscation is all you want, and make 
sure to translate them back (translating character entities never 
cascade, i.e., &32; stands for the string " " and not for the 
entity  , as such you need to do extra translation to go from 
  back to a space char)
 3. Use CDATA sections. But this creates different strings, the same 
way as with (2).
 4. Use string-to-codepoints() to get the codepoints (which map 
directly to the entities) of the string.

I think (2) and (3) will work well with XSLT 1 as well, but both require 
you to do extra work in your obfuscating tool. If you "just" want to 
have your text translated to character entities, without the serializer 
(correctly) making them into readable US-ASCII, option (1) is the 
easiest and most extensible to do.

Cheers,
-- Abel Braaksma