[Python-3000] Set literals - another try

[Josiah Carlson]

"Collin Winter" <collinw at gmail.com> wrote:
> 
> On 8/8/06, Josiah Carlson <jcarlson at uci.edu> wrote:
> > I personally don't see much of a use for set literals, considering that
> > there is a non-ambiguous spelling of it currently; set(...), whose only
> > cost above and beyond that of a set literal is a global name lookup.
> 
> I thought one of the main arguments in favor of set literals is that a
> literal form would allow the compiler to perform optimisations that
> the set(...) spelling doesn't allow.

The optimization argument used to define language syntax seems a bit
like the "tail wagging the dog" cliche.  For immutable literals that are
used a huge number of times (int, tuple, and other immutables), a
literal syntax for compiler optimization makes sense.  But for mutables
(list, dict, etc.), literal syntax is more a convenience as than an
optimization, as the compiler hasn't historically created once and
copied for re-use, but pushed values on the stack and called the
relevant create list bytecode. [1]

 - Josiah


[1]
Python 2.4.3 (#69, Mar 29 2006, 17:35:34) [MSC v.1310 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import dis
>>> def foo():
...     return [1,2,3]
...
>>> def goo():
...     return (1,2,3)
...
>>> dis.dis(foo)
  2           0 LOAD_CONST               1 (1)
              3 LOAD_CONST               2 (2)
              6 LOAD_CONST               3 (3)
              9 BUILD_LIST               3
             12 RETURN_VALUE
>>> dis.dis(goo)
  2           0 LOAD_CONST               4 ((1, 2, 3))
              3 RETURN_VALUE
>>>