Re: [SQLObject] Unicode support

[Ian B. <ia...@co...>]

On Nov 9, 2003, at 8:32 AM, John Baker wrote:
> Hello,
>
> Does SQLObject have any kind of Unicode support? We seem unable to 
> find any,
> and we've put a few hacks in there to encode Strings as utf-8 before 
> they are
> placed into the database, and decodes them after they come out. It's 
> nothing
> special, but I do feel SQLObject is lacking something here.

No, there's no specific Unicode support, mostly because I'm not sure 
how they should best be encoded -- if they should be UTF-8 encoded, if 
databases have unicode literals, etc.

If you wanted to make a validator/converter to do UTF-8 encoding, it 
might look like:

from SQLObject.include import Validator
class UTF8Validator(Validator.Validator):
     def fromPython(self, v, state):
         # test isinstance(v, unicode)?
         return v.encode('UTF-8') # strict, loose?
     def toPython(self, v, state):
         return v.decode('UTF-8')

You could then use something like:

class MyThing(SQLObject):
     unicodeColumn = String(validator=UTFValidator())

I'm not entirely sure about all the details, but this pattern could 
probably be added to Col.py, and show up either as a subclass of 
String, or something that could be enabled with a keyword argument.  
(In that case it probably would be good to include other encodings, 
especially for legacy databases that don't use UTF-8, and perhaps some 
clearer handling of strict vs. loose and how to handle non-unicode 
strings)

--
Ian Bicking | ia...@co... | http://blog.ianbicking.org