Yes, I realize that there is a work-around, but none of us should have to deal
with this mess in our implementations. Instead, we need to change the spec
so that the implementations will be simpler. The lack of ordering with regard
to charset and language has no value.
"ipp-attribute-fidelity" is not the same problem because its value applies to
the following group. That is, the ordering rules work in this case.
Bob Herriot
At 11:56 AM 3/27/98 , Carl Kugler wrote:
>I found it rather easy to implement when I did it.
>>"ipp-attribute-fidelity" is another special attribute because it, too,
>determines how all attributes should be interpreted.
>>I think there's no getting around saving some state during initial attribute
>processing which is later used to evaluate the operation as a whole.
>>Certainly you could save the raw form of any 'text' or 'name' strings
during a
>first pass over the attributes, then interpret them using the supplied
>"attributes-charset" and "attributes-natural-language" in a later phase
>