Source

pypy / pypy / objspace / std / settype.py

Author Commit Message Labels Comments Date
Philip Jenvey
add the rest of the needed __length_hint__ impls and more tests
Branches
length-hint
Lukas Diekmann
W_SetObject not takes w_iterable as init value instead of r_dict
Branches
set-strategies
Amaury Forgeot d'Arc
kill most unwrap_spec in objspace
Branches
improve-unwrap_spec
Dan Villiom Podlaski Christiansen
No-op merge to register the merge done in 0e3099e3cc33 with Mercurial.
Branches
fast-forward
David Schneider
merged trunk changes r77502 through r79385 into arm backend branch
Branches
arm-backend
Armin Rigo
set.difference_update(more_than_one_argument).
Branches
fast-forward
Armin Rigo
set.intersection_update(more_than_one_argument).
Branches
fast-forward
Armin Rigo
set.update(more_than_one_argument).
Branches
fast-forward
Armin Rigo
set.difference(more_than_one_argument).
Branches
fast-forward
Armin Rigo
set.intersection(more_than_one_argument).
Branches
fast-forward
Armin Rigo
set.union(more_than_one_argument).
Branches
fast-forward
Armin Rigo
Set's isdisjoint() method.
Branches
fast-forward
Amaury Forgeot d'Arc
Merge from trunk: -r78251:78315
Branches
fast-forward
Amaury Forgeot d'Arc
Set list.__hash__ is now None, instead of a function that raises TypeError. no_hash_descr() is gone! This fixes isinstance([], _abcoll.Hashable)
Branches
fast-forward
Armin Rigo
Added docstrings to all (non-special) methods of built-in types in the StdObjSpace.
Stephan Diehl
removed the WITHSET switch.
Armin Rigo
Just for fun, or maybe to try to win the largest diff award: refactored the StdObjSpace to no longer store the 'space' attribute on any of the W_XxxObject classes. It was not really used in any essential way any more.
Armin Rigo
Hash support: some Wrappables were missing a hash. The main work in this check-in, though, is about making sure the hash_cache field inserted by the rtyper for instances on which hash() is called, only shows up in selected subclasses of W_Root instead of on W_Root. Classes like W_IntObject and W_TupleObject don't need it. (This is a step towards tagged pointers, which has no chance to work if the W_SmallIntObject class contains more than an 'intval' field.)
Stephan Diehl
somewhat finished set/frozenset implementation. Needs some more debugging because translation fails in "backendopt:inlining". Checkin is for documentation purposes only.
Stephan Diehl
all tests in lib-python/modified-2.4.1/test/test_set.py are now passing, but the pickling for subtypes of set and frozenset (Actually, when running the commands in CLI, there is not problem). But, translation doesn't work yet. Still investigation (and fighting with the debugging process). As always, pypy.objspace.std.model.WITHSET must be set to True to enable native set/frozenset implementation.
Stephan Diehl
created extra 'frozensettype.py' file to avoid multimethod registering problems. The 'setobject.py' could stay more or less the same, but it had to be taken care that 'register_all' gets only the appropriate selection of methods for each type.
Stephan Diehl
somewhat cleaner set/frozenset implementation. set 'WITHSET = True' in model.py
Stephan Diehl
complex numbers are now considered working: tests are passed and pypy-c can be created. set implementation is nearly done. To test set/frozenset, enable WITHSET in objspace/std/model.py
Stephan Diehl
initial checkin of set/frozenset implementation for std objspace. Since there are no tests yet (yes, I know), the WITHSET switch needs to be set to 'True' in order to use native sets in pypy.