o Both encode_value and decode_value uses the master itself as fallback.
Ah, that makes a bit more sense then...
/ David Hedbor
Previous text:
2003-01-06 13:05: Subject: encode_value for objects
It seems to be a bit confused:
o The doc for decode_value says it uses the master itself as codec if none is given.
o The doc for encode_value says nothing but hints that a codec object must be given explicitly: "To encode objects, programs and functions, a codec object must be provided."
o Both encode_value and decode_value uses the master itself as fallback.
o The master doesn't implement the functions to act as a codec (the Codec class in the master does).
I'd like to solve this by removing the fallback to the master, thereby requiring a codec to be given explicitly. One reason is that it would avoid the often unnecessary overhead of instantiating a Codec object. Another is that more explicit control over the codec almost always is required to handle objects, functions and programs in a good way.
/ Martin Stjernholm, Roxen IS