[path-PEP] Path inherits from basestring again

[Reinhold Birkenfeld]

phil hunt wrote:
> On Sat, 30 Jul 2005 19:01:49 +0200, Reinhold Birkenfeld <reinhold-birkenfeld-nospam at wolke7.net> wrote:
>>phil hunt wrote:
>>
>>> def normalizePath(p, *pathParts):
>>>    """ Normalize a file path, by expanding the user name and getting 
>>>    the absolute path..
>>>    @param p [string] = a path to a file or directory
>>>    @param pathParts [list of string] = optional path parts
>>>    @return [string] = the same path, normalized
>>>    """
>>>    p1 = os.path.abspath(os.path.expanduser(p))
>>>    if len(pathParts)>0:
>>>       allPathParts = [ p1 ]
>>>       allPathParts.extend(pathParts)
>>>       p1 = os.path.join(*allPathParts)
>>>    p2 = os.path.abspath(p1)   
>>>    return p2
>>> normalisePath=normalizePath # alternate spelling 
>>> join=normalizePath # it works like os.path.join, but better  
>>> 
>>> 
>>> To be honest I don't see the point of having a Path class. That's 
>>> the way Java does it, and I find path handling in Java to be a lot 
>>> more of a hassle than in Python. (Actually, most things are more of 
>>> a hassle in Java, but that's another story).
>>
>>You see, with the Path class the above function could be written as
>>
>>def normalizePath(p, *pathParts):
>>    """ Normalize a file path, by expanding the user name and getting
>>    the absolute path..
>>    @param p [Path] = a path to a file or directory
>>    @param pathParts [list of string/Path] = optional path parts
>>    @return [Path] = the same path, normalized
>>    """
>>    tp = p.expanduser().abspath()
>>    return tp.joinwith(*pathParts).abspath()
>>
>>That's clearly an improvement, isn't it?
> 
> An improvement to what? To how the class is implemented, or to how 
> it is used?

No, the second function is cleaner and more readable than the first,
IMHO.

> If you mean the former, yes is it, due to the os.path module not 
> providing a function that does this. 
> 
> If you mean the latter, I disagree, because I would then have to 
> call it with something like:
> 
>    pn = normalizePath(Path(p), q)

That's easily helped by s/tp = p/tp = Path(p)/.

> and then I would have the problem that (pn) isn't a string so 
> calling a function to write some data into the file at that filename 
> would no longer work, i.e. this:
> 
>    writeFile(pn, someData)
> 
> would become this:
> 
>    writeFile(pn.getString(), someData)

Why? A Path is a string.

Reinhold