pydoc and latin-1/iso-8859-1/scandinavian letters

[Gerhard Häring]

Jorge Godoy wrote:
> Gerhard Häring <gh at ghaering.de> writes:
>>I'd recommend you switch to pyPgSQL. It supports Unicode and I've
>>heard of one of our users successfully using it in combination with
>>a Unicode-enabled wxPython.
> 
> Hmmm... It is possible to make that switch if this is the only
> solution to that problem

There should certainly be possible solutions or workarounds for 
PyGreSQL's pgdb.

> but then I'd have to create and maintain
> packages for that. 

The alternatives pyPgSQL and psycopg are already packaged quite well for 
Unix. For all *BSDs and Linux distributions I care about there are 
already packages.

> We have developers in Linux and Windows XP doing
> some collaborative work for a project here. 

pyPgSQL is also available as a Windows installer. In fact, I'm the guy 
who does the win32 stuff for pyPgSQL :-)

> pgdb was chosen because it is standard with PostgreSQL

That is is currently contained in the PostgreSQL downloads and CVS has 
probably more historical than technical reasons.

This will change. The database adapters (Python, Perl, Java, ...) are or 
are being removed from the PostgreSQL distribution. Most of them are 
being maintained at GBORG now.

> and available
> in both Linux distributions we have: Conectiva and Mandrake. It won't
> be hard to do that, but I would really like to know if there's a pgdb
> solution at all... 

There will certainly be. As long as you're using 8-bit strings and no 
Unicode strings, all you can do at the client side is to send a "SET 
CLIENT_ENCODING TO ...", anyway.

>>Maybe psycopg supports Unicode as well nowadays, I haven't checked.
>>
>>Anyway, you'll want to check the encodings your PostgreSQL client and
>>server uses (SET CLIENT_ENCODING TO ..., etc.). The PostgreSQL handbooks
>>explains that stuff pretty well IMO.
> 
> I've already done that, as I said in my prior message. [...]

Sorry, I didn't read it all too carefully it seems.

-- Gerhard