Commit Graph

37 Commits

Author SHA1 Message Date
Guido van Rossum
50e92235e7 Explicitly raise an exception in __cmp__ -- this clarifies that cmp()
is not supported on sets.  (Unfortunately, sorting a list of sets may
still return random results because it uses < exclusively, but for
sets that inly implements a partial ordering.  Oh well.)
2003-01-14 16:45:04 +00:00
Raymond Hettinger
35e48d2426 SF 643115: Set._update() had a special case for dictionaries which allowed
non-true values to leak in.  This threw-off equality testing which depends
on the underlying dictionaries having both the same keys and values.
2002-11-25 20:43:55 +00:00
Jeremy Hylton
cd58b8f532 Add getstate and setstate implementation to concrete set classes. 2002-11-13 19:34:26 +00:00
Guido van Rossum
7cd83ca9ad Another attempt at making the set constructor both safe and fast. [SF
bug 628246]
2002-11-08 17:03:36 +00:00
Tim Peters
0ec1ddcdcf _update(): Commented the new obscurity. Materialized into a tuple
instead of into a list for a bit of speed/space savings.  Reopened the
bug report too (628246), as I'm unclear on why we don't sort out the
cause of the TypeError instead.
2002-11-08 05:26:52 +00:00
Raymond Hettinger
1eb1fb814b Closes SF bug #628246.
The _update method detected mutable elements by trapping TypeErrors.
Unfortunately, this masked useful TypeErrors raised by the iterable
itself.  For cases where it is possible for an iterable to raise
a TypeError, the iterable is pre-converted to a list outside the
try/except so that any TypeErrors propagate through.
2002-11-08 05:03:21 +00:00
Raymond Hettinger
bfcdb8734e .iterkeys() is not needed. 2002-10-04 20:01:48 +00:00
Raymond Hettinger
1a8d193121 Sped _update().
Uses the fast update() method when a dictionary is available.
2002-08-29 15:13:50 +00:00
Tim Peters
454602f0f7 Gave intersection_update a speed boost. 2002-08-26 00:44:07 +00:00
Tim Peters
cd06eeb20c Gave issubet() and issuperset() major speed boosts. That's it for now!
Someone else may want to tackle the mutating operations similarly.
2002-08-25 20:12:19 +00:00
Tim Peters
b8940393e9 Gave __sub__/difference a factor of 2-5 speed boost. 2002-08-25 19:50:43 +00:00
Tim Peters
334b4a5c39 Gave __xor__/symmetric_difference a factor of 2-5 speed boost. 2002-08-25 19:47:54 +00:00
Tim Peters
37faed2532 Sped union by a factor of 3-4. 2002-08-25 19:21:27 +00:00
Tim Peters
d33e6be59d Sped intersection by large factors (3-5x faster than before on sets of
cardinality 500; and the smaller the intersection, the bigger the speedup).
2002-08-25 19:12:45 +00:00
Tim Peters
4a2f91e302 Added a clue about why xyz_update isn't the same as __xyz__. 2002-08-25 18:59:04 +00:00
Tim Peters
ea76c98014 Implemented <, <=, >, >= for sets, giving subset and proper-subset
meanings.  I did not add new, e.g., ispropersubset() methods; we're
going nuts on those, and, e.g., there was no "friendly name" for
== either.
2002-08-25 18:43:10 +00:00
Tim Peters
4924db176b Record a clue about why __or__ is not union, etc. 2002-08-25 17:10:17 +00:00
Raymond Hettinger
e87ab3fefe Removed < <= > >= from the API. Implemented as comparisons of the
underlying dictionaries, there were no reasonable use cases (lexicographic
sorting of a list of sets is somewhat esoteric).  Frees the operators
for other uses (such as strict subset and superset comparisons).

Updated documentation and test suite accordingly.
2002-08-24 07:33:06 +00:00
Raymond Hettinger
1b9f5d4c1a At Tim Peter's suggestion, propagated GvR's binary operator changes to
the inplace operators.  The strategy is to have the operator overloading
code do the work and then to define equivalent method calls which rely on
the operators.  The changes facilitate proper application of TypeError
and NonImplementedErrors.

Added corresponding tests to the test suite to make sure both the operator
and method call versions get exercised.

Add missing tests for difference_update().
2002-08-24 06:19:02 +00:00
Raymond Hettinger
d50185127f Since instances of _TemporarilyImmutableSet are always thrown away
immediately after the comparison, there in no use in caching the hashcode.
The test, 'if self._hashcode is None', never fails.  Removing the caching
saves a few lines and a little time.
2002-08-24 04:47:42 +00:00
Raymond Hettinger
fa1480f686 1. Removed module self test in favor of unittests -- Timbot's suggestion.
2. Replaced calls to Set([]) with Set() -- Timbot's suggestion
3. Fixed subtle bug in sets of sets:

The following code did not work (will add to test suite):
    d = Set('d')
    s = Set([d])  # Stores inner set as an ImmutableSet
    s.remove(d)   # For comparison, wraps d in _TemporarilyImmutableSet

The comparison proceeds by computing the hash of the
_TemporarilyImmutableSet and finding it in the dictionary.
It then verifies equality by calling ImmutableSet.__eq__()
and crashes from the binary sanity check.

The problem is that the code assumed equality would be checked
with _TemporarilyImmutableSet.__eq__().

The solution is to let _TemporarilyImmutableSet derive from BaseSet
so it will pass the sanity check and then to provide it with the
._data element from the wrapped set so that ImmutableSet.__eq__()
will find ._data where it expects.

Since ._data is now provided and because BaseSet is the base class,
_TemporarilyImmutableSet no longer needs .__eq__() or .__ne__().

Note that inheriting all of BaseSet's methods is harmless because
none of those methods (except ones starting with an underscore)
can mutate the .data element.  Also _TemporarilyImmutableSet is only
used internally as is not otherwise visible.
2002-08-24 02:35:48 +00:00
Tim Peters
53506be258 pop() docstring: this isn't a randomly-chosen element, it's merely
arbitrary.  I already changed the docs for this.
2002-08-23 20:36:58 +00:00
Tim Peters
d06d03041b Comment repair. 2002-08-23 20:06:42 +00:00
Guido van Rossum
e399d08a4a RH pointed out that discard(element) doesn't do the transformation on
the element if necessary.  Fixed by calling self.remove(element).
2002-08-23 14:45:02 +00:00
Guido van Rossum
dc61cdf6c0 Change the binary operators |, &, ^, - to return NotImplemented rather
than raising TypeError when the other argument is not a BaseSet.  This
made it necessary to separate the implementation of e.g. __or__ from
the union method; the latter should not return NotImplemented but
raise TypeError.  This is accomplished by making union(self, other)
return self|other, etc.; Python's binary operator machinery will raise
TypeError.

The idea behind this change is to allow other set implementations with
an incompatible internal structure; these can provide union (etc.) with
standard sets by implementing __ror__ etc.

I wish I could do this for comparisons too, but the default comparison
implementation allows comparing anything to anything else (returning
false); we don't want that (at least the test suite makes sure
e.g. Set()==42 raises TypeError).  That's probably fine; otherwise
other set implementations would be constrained to implementing a hash
that's compatible with ours.
2002-08-22 17:23:33 +00:00