On Saturday, September 7, 2002, at 10:03 AM, Karl Ove Hufthammer wrote:

> What is meant by the '2002-09-04 17:46:13' commit for the localisation
> strings files: 'the stringset is now encoding aware'
>
> For example:
>
> <AbiStrings app="AbiWord" ver="1.0" language="nn-NO">
>
> was changed to:
>
> <AbiStrings app="AbiWord" ver="1.0" language="nn-NO" 
> encoding="iso-8859-1">
>
> XML files *already* contains encoding information in the XML
> declaration:
>
> <?xml version="1.0" encoding="iso-8859-1"?>
>
> and it's *impossible* to use a different encoding for elements
> than the one used for the entire XML file (i.e. the XML file
> wouldn't be well-formed, and no XML parser would able to read it).
> This change just doesn't make sense.

You tell me how to get the document's encoding without having the 
encoding inside of and element's tag (and no, <?xml?> is not an element 
  and never goes through any callbacks) for expat, libxml2, and MSXML, I 
give you a cookie and this goes away. Ideally you'd add a virtual 
method to UT_Xml::Reader and then implement it for the 3 subclasses. 
Until then, it stays, as it was the easiest solution to the problem and 
causes no harm whatsoever except ~20 bytes of duplicated data. 
"Whoop-dee-doo."

Dom

Reply via email to