[Pythonmac-SIG] Re:mac islink()

[Chris Barker]

Jack Jansen wrote:
>Ah, memories.... (That comment is still from the time that Guido
>himself was responsible for MacPython)
>
>But: I'm not 100% sure that an alias is sufficiently similar to a
>symlink that it would be a good idea to put this in. Hmm, let me check 
>where islink is used... Hmm, inconclusive results. Some scripts would
>definitely break, most appear to be unix-specific anyway. One other
>point against fixing islink is that ntpath.islink() also always
>returns zero, while NT aliases are a lot closer to unix symlinks than
>MacOS aliases (if I understand them correctly).

>I think I'll just take votes. So let me know whether you want islink() 
>to remain as-is or fixed.

I vote to keep it as it is. Platform dependent stuff has no place in the
os module with the same name, and alias' are NOT the same as sym-links.

Frankly, I don't think ANYTHING that is platform dependent should be in
the os module. I don't know what is being planned for Py3k, but I'd love
to see everything taken out of os that isn't cross platform. That way,
I'd know as soon as I wote the code that it would work on all platforms.
If I have to import macos, I'll know it'll only work on a mac.

-Chris

-- 
Christopher Barker,
Ph.D.                                                           
cbarker@jps.net                      ---           ---           ---
http://www.jps.net/cbarker          -----@@       -----@@       -----@@
                                   ------@@@     ------@@@     ------@@@
Water Resources Engineering       ------   @    ------   @   ------   @
Coastal and Fluvial Hydrodynamics -------      ---------     --------    
------------------------------------------------------------------------
------------------------------------------------------------------------