Re: [Pyobjc-dev] non-default encoding raising an exception over the bridge

[Bob I. <bo...@re...>]

On Nov 22, 2003, at 12:08 AM, Ronald Oussoren wrote:

>
> On 22 nov 2003, at 4:51, Bob Ippolito wrote:
>
>> I think that most of the time, people would be happier if this didn't 
>> raise an exception (use the unicode replace instead).  If the python 
>> user doesn't know the encoding of the text, and it's not default 
>> encoding, the developer probably doesn't have anything better than 
>> replace to use on the string.  The exception is a nuisance and can be 
>> hard to track down.
>>
>> Perhaps this could be a module level variable in objc, so that a 
>> person could toggle it to be strict if they wanted it that way, but 
>> have the default as replace?
>>
>> -bob
>
> We already have such a variable, it is called defaultencoding and can 
> be set using sys.setdefaultencoding().

only at the "sitecustomize" phase.. you can't change this later.  I 
don't really care to change the encoding, just what happens if it can't 
encode.

> This is not exactly what you ask, but IMHO better than what you 
> propose ("Errors should never pass silently.").
>
> BTW. You should convert all input to unicode instead of waiting for 
> problems with the implicit conversion to unicode that is performed by 
> PyObjC. You're more likely to know the right encoding while reading 
> the data.

But what's better, seeing unicode replace characters in your app or 
trying to track down a strange exception?  If the behavior stays the 
same, maybe it should give you a hint as to what character sequence 
gave you the error.

-bob