[tahoe-dev] #534: "tahoe cp" command encoding issue

[Shawn Willden]

Yeah, I understand the limitations of JSON; I'm using it for my backup code.

My comment was just that Base32/Base64 aren't the only encoding options.  For 
filenames, I think the "assume latin1" is a better choice.   For the western 
world, at least, the result is likely to be more compact and will usually be 
somewhat readable when viewed as ASCII, which is convenient.  For the 
non-Western world I suspect that if you have any but the right encoding the 
result will be gibberish.

I was using Python escaping to handle uninterpretable filenames, but I've 
switched to latin1 and I think it's much cleaner, even if it is arguably an 
abuse of the codec.

	Shawn.