Is UserList.py "strange" or is it cmp()?

[Tom Funk]

Hi, Alex.

Thanks for the response. 

In an article posted 14 Mar 2000 14:37:49 -0500,
Alex (alex at somewhere.round.here) said:
> Usually, the way it's supposed to work is
> 
> cmp (a, b) => -1  if a < b
> cmp (a, b) =>  0  if a == b
> cmp (a, b) =>  1  if a > b

True, consistent with the C library's strcmp() family.
> 
> which is consistent with the results you got.  I guess that 'a == b' is
> actually returning (not cmp (a, b))
> 
> Alex.

Duh! Epiphany! I just got it! 

cmp() and __cmp__() both function the same way: returning negative, zero 
or positive values when comparing two objects -- cmp() is intended for 
ORDERING items.  The == operator takes the cmp() result and indicates 
true (1) or false (0) if the two items are are equal (I.e., cmp(a,b)==0).  
If we were to find the runtime function that handles "==", it might look 
like this:

   return cmp(pyObject1, pyObject2) == 0;

As you suggested, using:

   return (not cmp(pyObject1, pyObject2));

would yield the same result.

But that doesn't tell us why testuserlist.py dies at this assertion:

  assert u2[:I] == l2[:I]

Seems to me this assertion should pass.

-- 
-=< tom >=-
Thomas D. Funk                           |       "Software is the lever
Software Engineering Consultant          | Archimedes was searching for"
Advanced Systems Design, Tallahassee FL. |