[RFC][#30159] Fix to make bzr st and bzr st -S display paths relative to cwd.

Ian Clatworthy ian.clatworthy at internode.on.net
Tue Feb 10 02:30:42 GMT 2009


Matthew D. Fuller wrote:
> On Mon, Feb 09, 2009 at 08:14:17AM -0800 I heard the voice of
> Eddie Parker, and lo! it spake thus:
>> Ah, yes.  Regarding the 'spurious whitepaces', [...]
> 
> Well, *I*'m big on lots of whitespace.  But the bzr devs aren't so
> much.

Like any bunch of developers, I suspect opinions vary w.r.t. whitespace
across the bzr core devs. We agree on PEP-8 at least. :-)

In my case at least, I've long been a believer in most of Code Complete
(http://cc2e.com/Page.aspx?nid=71) including the basic stuff like the
so-called "Pseudocode Programming Process". That tends to lead to code
like ...

  # blah blah blah
  ...

  # blah blah blah
  ...

Works for me. Even with syntax highlighting editors, I find the blank
lines between blocks of logic valuable.

>> May I ask why?  I'm honestly curious, as most other VCS I use don't
>> have this concept of printing from the root and I've never found
>> myself wanting it
> 
> Because I have a mental model of my project[s], and with the full
> paths I always know what files it's talking about.  With cwd-relative
> paths, I have to know where I am in the tree and then mentally sort
> out arbitrarily high ascensions to figure out what's actually changed.
> 
> I note that 'hg stat' prints root-relative paths.  So does 'mtn stat'.
> Sure, svn and CVS are cwd-relative, but they only show cwd-descended
> status anyway.

Right. We probably need to do a clearer job of explaining the whole
tree-centric thing in the User Guide, and ensuring we're consistent
across commands. Having a --from-root option across the (core at least)
commands would be nice and could be introduced in the doc there.

Ian C.



More information about the bazaar mailing list