tla changes vs. baz status

[John Whitley]

On Jun 21, 2005, at 6:59 AM, Martin Pool wrote:
>
> More seriously, the current default is meant to be a tradeoff between
> being machine-parseable and being reasonably understandable.  I used
> to like 'svn status', but I think Aaron's right that having to refer
> to the legend so often shows something is wrong.

Here's a question: why tradeoff at all?  Have "bzr status" default to  
a human-readable format that (in principle) may change from release  
to release to ease human understanding issues.  Think of its output  
like compiler error messages: designed to aid the human user.

Complement that with something like "bzr status --xml" that outputs a  
cleanly spec'd machine-readable format.  With XML as the output, no  
one will have to mess about reinventing a parser for it in the first  
place.  If an XML schema is available, then output from "bzr status -- 
xml" can be run through a validator as part of the test suite.

-- John