Re: [PATCH] Charset generalization - take 1

	  Paul,

> There are still things left to test. For example, when I pass encoding
> to xmlReadMemory and xml source have <?xml> preamble with different
> encoding specified, which will prevail? For our cause, it will be
> better if the former will be preferred. But in reality I need to
> consult libxml2 source code :)

      Yeah, I'd wondered about that.  The fact that you pass in a 
charset makes me think it'd prevail over an encoding=, but you never 
know.  I will say that with the xmlParseMemory() call used before, the 
passed in charset did win out.  For my fix I'd pass in utf-8 with an 
encoding= of iso-8859-2 and come out with the compiled form saying 
utf-8.  I never had to strip or change the encoding= block to make it work.

> I have this thing online for testing and it works for some sites with
> my phone (nokia 6310i), but for extensive testing I'll try to put it
> online to production :)

      You're a very brave man.  :P  I have been doing my testing with 
either the Ericsson WapIDE or via a carrier (both on a devel box) using 
the Motorola RAZR.  I'll do the same testing again as I am sure the 
above possible issue is a non-issue.

> Don't worry,I'll put it somewhere else :)

      Hide an obscure leak in one of the octstr functions.  That'd kill 
a box quick.  :P

Jon