user_encoding fix

[Robey Pointer]

On 16 Feb 2006, at 16:27, Nir Soffer wrote:

> I branched John A. Meinel encoding branch, and started with a small  
> fix for user_encoding, described in http://bazaar.canonical.com/ 
> DarwinCommandLineArgumentDecoding.
>
> See revisions 1575..1577 from http://nirs.dyndns.org/bzr/encoding-nirs
>
> There are 3 new tests, 2 are trivial tests that check the correct  
> value, the third is a test that try to validate the value of  
> user_encoding by using shell completion and trying to decode the  
> completed names, which is relevant to all posix like platforms.
>
> Please comment and try the tests on your platform. I use here Mac  
> OS X 10.3, and have also a W2K machine but try to avoid it.

My terminals are set to use UTF8 encoding, but I think that was  
because I set them myself.  I believe Macs in the US locale may use a  
default encoding of Latin-1.

If that's not true, and Macs always use a default terminal encoding  
of UTF8, then I'm +1 on just forcing darwin/osx to always use utf8  
encodings inside bzr.

robey