Re: [cp-patches] gnu/xml/transform/StreamSerializer.java: compatibilityMode setting

[Chris Burdess]

Ito Kazumitsu wrote:
> gnu/xml/transform/StreamSerializer.java sets compatibilityMode to true
> for any encoding other than UTF* or UCS*.
>
> This is very inconvenient for people using non-ASCII encodings,
> especially CJK encodings, where most of the characters > 127.
>
> compatibilityMode being set to true, the output grows too large
> and becomes completely human-unreadable.
>
> I would like by some means to set compatibilityMode to false,
> and this is what I did.
>
> --- gnu/xml/transform/StreamSerializer.java.orig	Fri Dec 24 07:38:44 
> 2004
> +++ gnu/xml/transform/StreamSerializer.java     Sun Feb 13 08:43:33 2005
> @@ -106,6 +106,12 @@
>              compatibilityMode = false;
>            }
>        }
> +    String sysprop = System.getProperty(
> +        "gnu.xml.transform.StreamSerializer.compatibilityMode");
> +    if (sysprop != null)
> +      {
> +        compatibilityMode = ("true".equalsIgnoreCase(sysprop));
> +      }
>      this.eol = (eol != null) ? eol : 
> System.getProperty("line.separator");
>      namespaces = new HashMap();
>    }

Unfortunately your patch is almost guaranteed to produce 
non-well-formed XML.

I agree that compatibilityMode is a hack. What's really needed is a way 
to detect whether a character is a valid member of a given encoding, 
and what its value in that encoding is. If it isn't a valid member of 
that encoding, it needs to be escaped via the character entity 
reference mechanism. We can't simply use String.getBytes(encoding), as 
this will simply lose the characters that are not valid members of the 
encoding (converting them to ASCII 63 iirc).
--
Chris Burdess