[tahoe-dev] Base32

[zooko]

Hello, Jeremy Fishman.  Thanks for looking at the tahoe source.

>   I personally agree with the use of lowercase, but would this be  
> better accomplished via base64.b32encode(...).lower()

There's one more difference, which is visible only when the length of  
the input in bytes isn't a multiple of 5:

 >>> for l in range(1, 7):
...  print base64.b32encode('x'*l).lower(), allmydata.util.base32.b2a 
('x'*l)
...
pa====== pa
pb4a==== pb4a
pb4hq=== pb4hq
pb4hq6a= pb4hq6a
pb4hq6dy pb4hq6dy
pb4hq6dypa====== pb4hq6dypa

I figured that it would be roughly as easy to re-use my own zbase32  
implementation (which I wrote and therefore like) as to strip the  
padding from the output of base64.

I would not reject a patch to replace zbase32 with suitably wrapped  
base64, especially if it came with tests of the padding situation and  
measurements of performance.  :-)

Regards,

Zooko