Commits

Author Commit Message Labels Comments Date
Armin Rigo
A different attempt: change jit/backend/x86. If stacklets are actually used (as detected by a quasi-immutable field read), we generate backend code that uses malloc()/realloc()/free() instead of storing stuff in the stack. This might solve most of the problems that we get so far by accessing 'vable_token' and 'virtual_token': they would become heap pointers, and so remain valid even if the stack is moved away. The performance impact should be null as long as we don't use stacklets, but when we do, we have to consider: * calls to malloc()/realloc()/free() whose result goes into 'ebp'; maybe later we can use a custom version of these functions, particularly for realloc() at the start of every bridge. * on the other hand, it makes the stack much smaller if everything is JITted, which is g…
Branches
continulet-jit-2
Armin Rigo
Random progress.
Branches
continulet-jit
Armin Rigo
Adapt tests.c.
Branches
continulet-jit
Armin Rigo
Change again the API: this (unimplemented) version looks like it can be used by the JIT.
Branches
continulet-jit
Armin Rigo
hg merge default
Branches
continulet-jit
Armin Rigo
Clarify.
Branches
continulet-jit
Antonio Cuni
refactor hex() and oct(). In py2 they support the special methods __hex__ and __oct__, and thus are rendered as regular objspace methods (space.oct and space.hex). In py3k, __hex__ and __oct__ are gone, so they are just normal builtins, which are implemented on top of __format__ for consistency with bin() (the alternative would be to implement them on top of the rpython hex() and oct()). Killing support for hex and oct f…
Branches
py3k
Antonio Cuni
make sure that bin() calls __index__
Branches
py3k
Antonio Cuni
kill the config docs for two old modules
Branches
py3k
Armin Rigo
Slight changes of the interface, to make it clear that callers don't expect to do anything with the token --- just check if it's 0 or not.
Hakan Ardo
hg merge
Hakan Ardo
Dont import boxes proven constant while setting up the short_boxes and dont use the fallback to produce boxes with same_as if the box was already produced (should fix issue1045)
Hakan Ardo
simplify test
Maciej Fijalkowski
work some on the tests, IN PROGRESS
Branches
dead-code-optimization
Antonio Cuni
fix syntax
Branches
py3k
Antonio Cuni
kill the cases about longs. Also, instances of classes which define __eq__ but not __hash__ are unhashable in py3k. Fix the test, which now passes with -A but fails on py.py
Branches
py3k
Antonio Cuni
kill the case about longs, we no longer have them
Branches
py3k
Antonio Cuni
objects of incompatible types are now unordeable in python3. Completely change the meaning of this test: instead of checking the ordering, we check that we always raise TypeError. Passes with -A, still fails on pypy
Branches
py3k
Antonio Cuni
__hex__ and __oct__ has gone in py3k, and we can return only str from __str__ and __repr__ (instead of e.g. str and unicode in py2). Adapt this test, which is now very simple
Branches
py3k
Antonio Cuni
kill the u''
Branches
py3k
Antonio Cuni
kill the u''
Branches
py3k
Antonio Cuni
this syntax is no longer valid, kill the test
Branches
py3k
Antonio Cuni
kill the cStringIO module and its tests
Branches
py3k
Antonio Cuni
bah, the name of this class is clearly wrong (maybe a copy&paste)?
Antonio Cuni
kill the u prefix, and adapt che expected bytecode to the new py3k compiler
Branches
py3k
Hakan Ardo
failing test
Antonio Cuni
fix syntax
Branches
py3k
Antonio Cuni
remove a u'', s/unichr/chr/, kill 'long'
Branches
py3k
Antonio Cuni
test bytes literals instead of unicode, and make sure to write bytes to a file opened with 'wb'
Branches
py3k
Antonio Cuni
s/func_code/__code__
Branches
py3k
  1. Prev
  2. Next