I've modified redis_cache.py to be able to optionally provide an application name to RedisCache constructor. This makes possible to share cache between different web2py applications if needed.
I'm not sure if this would help someone else. I've always had this scenario where some of my web2py applications share the models, and therefore, need to share the cache (I use Redis). In the past I've asked if there was any possibility to share cache between applications, but I got no answers:
https://groups.google.com/forum/#!searchin/web2py/share$20cache$20application%7Csort:date/web2py/iNCzMq8ADnw/ufyPBWajBQAJ
Anyway, I've noticed that I could make a simple change to redis_cache.py to achieve what I was looking for: I've added the optional argument "application" to RedisCache(). If you inspect the RedisCache() code, you will notice that previously it was forced to use current.request.application for the instance_name. With this simple change I did, it's possible to provide an application name to the RedisCache() constructor.
This has worked like a charm for my specific scenario, and I think I can say that it has backwards compatibility. Still, if this proposal isn't accepted, I would like to hear about some alternative to share cache between applications. Thanks!
This lets an application use a custom JSONEncoder when making jsonrpc calls (to automatically handle, say, serialization of datetime, or Decimal, or custom classes).
e.g.:
import json, datetime, decimal
class ExtendedEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj, datetime.datetime):
return obj.isoformat()
if isinstance(obj, decimal.Decimal):
return str(obj)
return super(ExtendedEncoder, self).default(obj)
api = ServerProxy(address, version='2.0', encoder=ExtendedEncoder)
Currently, JSONRPCErrors get spit into a console or ticket one line per character (due to the '\n'.join(data) applying to a string). This fixes that issue while preserving new lines for arrays.
mdipierro