2010-01-11 11:14:40 Fabian Groffen napisał(a):
> On 11-01-2010 08:29:32 +0000, Duncan wrote:
> > Fabian Groffen posted on Mon, 11 Jan 2010 08:50:30 +0100 as excerpted:
> > 
> > > On 11-01-2010 01:25:45 +0100, Arfrever Frehtes Taifersar Arahesis wrote:
> > >> > Can you explain the intended use of this variable, and why normal
> > >> > DEPEND is not sufficient?
> > >> 
> > >> PYTHON_DEPEND is intented to simplify specification of dependency on
> > >> Python.
> > >> 
> > >> PYTHON_DEPEND="2:2.5" is shorter than: DEPEND="|| (
> > >> =dev-lang/python-2.7* =dev-lang/python-2.6* =dev-lang/python-2.5* )"
> > > 
> > > So if there is enough space to express the dependency with the current
> > > syntax, is it worth introducing a new shorthand for it then?
> > > 
> > > Also for this example, why does 2:2.5 expand to 2.7, 2.6 and 2.5?  I
> > > would have expected 2.0 ... 2.5.  Maybe the language isn't as intuitive
> > > then as Sebastian pointed out.
> > 
> > Initially intuitive, perhaps not, but reasonably easy after reading the 
> > explanation:
> > 
> > The first position is major python version, the second if present, 
> > minimal version (within that major), so it can be read as =>version, the 
> > third if present, maximum version (within that major), so it can be read 
> > as <=.
> > 
> > Thus, the above "2:2.5" means major version 2, minimal version 2.5 (no 
> > maximum version within that major), so 2.5+.
> 
> what's wrong with ">=2.5 && <3.0" then?

It's too long and harder to parse.

-- 
Arfrever Frehtes Taifersar Arahesis