[pypy-commit] pypy py3k: hg merge default
antocuni
noreply at buildbot.pypy.org
Wed Jul 18 16:19:05 CEST 2012
Author: Antonio Cuni <anto.cuni at gmail.com>
Branch: py3k
Changeset: r56162:47325d047a97
Date: 2012-07-18 16:15 +0200
http://bitbucket.org/pypy/pypy/changeset/47325d047a97/
Log: hg merge default
diff --git a/pypy/annotation/binaryop.py b/pypy/annotation/binaryop.py
--- a/pypy/annotation/binaryop.py
+++ b/pypy/annotation/binaryop.py
@@ -7,7 +7,7 @@
from pypy.tool.pairtype import pair, pairtype
from pypy.annotation.model import SomeObject, SomeInteger, SomeBool, s_Bool
from pypy.annotation.model import SomeString, SomeChar, SomeList, SomeDict
-from pypy.annotation.model import SomeUnicodeCodePoint
+from pypy.annotation.model import SomeUnicodeCodePoint, SomeStringOrUnicode
from pypy.annotation.model import SomeTuple, SomeImpossibleValue, s_ImpossibleValue
from pypy.annotation.model import SomeInstance, SomeBuiltin, SomeIterator
from pypy.annotation.model import SomePBC, SomeFloat, s_None
@@ -471,30 +471,37 @@
"string formatting mixing strings and unicode not supported")
-class __extend__(pairtype(SomeString, SomeTuple)):
- def mod((str, s_tuple)):
+class __extend__(pairtype(SomeString, SomeTuple),
+ pairtype(SomeUnicodeString, SomeTuple)):
+ def mod((s_string, s_tuple)):
+ is_string = isinstance(s_string, SomeString)
+ is_unicode = isinstance(s_string, SomeUnicodeString)
+ assert is_string or is_unicode
for s_item in s_tuple.items:
- if isinstance(s_item, (SomeUnicodeCodePoint, SomeUnicodeString)):
+ if (is_unicode and isinstance(s_item, (SomeChar, SomeString)) or
+ is_string and isinstance(s_item, (SomeUnicodeCodePoint,
+ SomeUnicodeString))):
raise NotImplementedError(
"string formatting mixing strings and unicode not supported")
- getbookkeeper().count('strformat', str, s_tuple)
- no_nul = str.no_nul
+ getbookkeeper().count('strformat', s_string, s_tuple)
+ no_nul = s_string.no_nul
for s_item in s_tuple.items:
if isinstance(s_item, SomeFloat):
pass # or s_item is a subclass, like SomeInteger
- elif isinstance(s_item, SomeString) and s_item.no_nul:
+ elif isinstance(s_item, SomeStringOrUnicode) and s_item.no_nul:
pass
else:
no_nul = False
break
- return SomeString(no_nul=no_nul)
+ return s_string.__class__(no_nul=no_nul)
-class __extend__(pairtype(SomeString, SomeObject)):
+class __extend__(pairtype(SomeString, SomeObject),
+ pairtype(SomeUnicodeString, SomeObject)):
- def mod((str, args)):
- getbookkeeper().count('strformat', str, args)
- return SomeString()
+ def mod((s_string, args)):
+ getbookkeeper().count('strformat', s_string, args)
+ return s_string.__class__()
class __extend__(pairtype(SomeFloat, SomeFloat)):
diff --git a/pypy/annotation/test/test_annrpython.py b/pypy/annotation/test/test_annrpython.py
--- a/pypy/annotation/test/test_annrpython.py
+++ b/pypy/annotation/test/test_annrpython.py
@@ -3389,6 +3389,22 @@
s = a.build_types(f, [str])
assert isinstance(s, annmodel.SomeString)
+ def test_unicodeformatting(self):
+ def f(x):
+ return u'%s' % x
+
+ a = self.RPythonAnnotator()
+ s = a.build_types(f, [unicode])
+ assert isinstance(s, annmodel.SomeUnicodeString)
+
+ def test_unicodeformatting_tuple(self):
+ def f(x):
+ return u'%s' % (x,)
+
+ a = self.RPythonAnnotator()
+ s = a.build_types(f, [unicode])
+ assert isinstance(s, annmodel.SomeUnicodeString)
+
def test_negative_slice(self):
def f(s, e):
diff --git a/pypy/doc/coding-guide.rst b/pypy/doc/coding-guide.rst
--- a/pypy/doc/coding-guide.rst
+++ b/pypy/doc/coding-guide.rst
@@ -255,7 +255,12 @@
code if the translator can prove that they are non-negative. When
slicing a string it is necessary to prove that the slice start and
stop indexes are non-negative. There is no implicit str-to-unicode cast
- anywhere.
+ anywhere. Simple string formatting using the ``%`` operator works, as long
+ as the format string is known at translation time; the only supported
+ formatting specifiers are ``%s``, ``%d``, ``%x``, ``%o``, ``%f``, plus
+ ``%r`` but only for user-defined instances. Modifiers such as conversion
+ flags, precision, length etc. are not supported. Moreover, it is forbidden
+ to mix unicode and strings when formatting.
**tuples**
diff --git a/pypy/objspace/std/strutil.py b/pypy/objspace/std/strutil.py
--- a/pypy/objspace/std/strutil.py
+++ b/pypy/objspace/std/strutil.py
@@ -193,4 +193,4 @@
try:
return rstring_to_float(mystring)
except ValueError:
- raise ParseStringError(u"invalid literal for float()")
+ raise ParseStringError(u"invalid literal for float(): '%s'" % s)
diff --git a/pypy/objspace/std/test/test_floatobject.py b/pypy/objspace/std/test/test_floatobject.py
--- a/pypy/objspace/std/test/test_floatobject.py
+++ b/pypy/objspace/std/test/test_floatobject.py
@@ -441,6 +441,14 @@
b = A(5).real
assert type(b) is float
+ def test_invalid_literal_message(self):
+ try:
+ float('abcdef')
+ except ValueError as e:
+ assert 'abcdef' in e.message
+ else:
+ assert False, 'did not raise'
+
def test_float_from_unicode(self):
s = '\U0001D7CF\U0001D7CE.4' # 𝟏𝟎.4
assert float(s) == 10.4
diff --git a/pypy/rlib/objectmodel.py b/pypy/rlib/objectmodel.py
--- a/pypy/rlib/objectmodel.py
+++ b/pypy/rlib/objectmodel.py
@@ -112,7 +112,9 @@
""" Decorate a function with forcing of RPython-level types on arguments.
None means no enforcing.
- XXX shouldn't we also add asserts in function body?
+ When not translated, the type of the actual arguments are checked against
+ the enforced types every time the function is called. You can disable the
+ typechecking by passing ``typecheck=False`` to @enforceargs.
"""
typecheck = kwds.pop('typecheck', True)
if kwds:
diff --git a/pypy/rlib/test/test_objectmodel.py b/pypy/rlib/test/test_objectmodel.py
--- a/pypy/rlib/test/test_objectmodel.py
+++ b/pypy/rlib/test/test_objectmodel.py
@@ -429,6 +429,7 @@
exc = py.test.raises(TypeError, "f(1, 2, 3)")
assert exc.value.message == "f argument number 2 must be of type <type 'str'>"
py.test.raises(TypeError, "f('hello', 'world', 3)")
+
def test_enforceargs_defaults():
@enforceargs(int, int)
diff --git a/pypy/rpython/lltypesystem/rstr.py b/pypy/rpython/lltypesystem/rstr.py
--- a/pypy/rpython/lltypesystem/rstr.py
+++ b/pypy/rpython/lltypesystem/rstr.py
@@ -1,5 +1,6 @@
from weakref import WeakValueDictionary
from pypy.tool.pairtype import pairtype
+from pypy.annotation import model as annmodel
from pypy.rpython.error import TyperError
from pypy.rlib.objectmodel import malloc_zero_filled, we_are_translated
from pypy.rlib.objectmodel import _hash_string, enforceargs
@@ -169,6 +170,13 @@
return result
@jit.elidable
+ def ll_unicode(self, s):
+ if s:
+ return s
+ else:
+ return self.convert_const(u'None')
+
+ @jit.elidable
def ll_encode_latin1(self, s):
length = len(s.chars)
result = mallocstr(length)
@@ -962,13 +970,18 @@
def do_stringformat(cls, hop, sourcevarsrepr):
s_str = hop.args_s[0]
assert s_str.is_constant()
+ is_unicode = isinstance(s_str, annmodel.SomeUnicodeString)
+ if is_unicode:
+ TEMPBUF = TEMP_UNICODE
+ else:
+ TEMPBUF = TEMP
s = s_str.const
things = cls.parse_fmt_string(s)
size = inputconst(Signed, len(things)) # could be unsigned?
- cTEMP = inputconst(Void, TEMP)
+ cTEMP = inputconst(Void, TEMPBUF)
cflags = inputconst(Void, {'flavor': 'gc'})
vtemp = hop.genop("malloc_varsize", [cTEMP, cflags, size],
- resulttype=Ptr(TEMP))
+ resulttype=Ptr(TEMPBUF))
argsiter = iter(sourcevarsrepr)
@@ -979,7 +992,13 @@
vitem, r_arg = argsiter.next()
if not hasattr(r_arg, 'll_str'):
raise TyperError("ll_str unsupported for: %r" % r_arg)
- if code == 's' or (code == 'r' and isinstance(r_arg, InstanceRepr)):
+ if code == 's':
+ if is_unicode:
+ # only UniCharRepr and UnicodeRepr has it so far
+ vchunk = hop.gendirectcall(r_arg.ll_unicode, vitem)
+ else:
+ vchunk = hop.gendirectcall(r_arg.ll_str, vitem)
+ elif code == 'r' and isinstance(r_arg, InstanceRepr):
vchunk = hop.gendirectcall(r_arg.ll_str, vitem)
elif code == 'd':
assert isinstance(r_arg, IntegerRepr)
@@ -999,9 +1018,17 @@
else:
raise TyperError, "%%%s is not RPython" % (code, )
else:
- from pypy.rpython.lltypesystem.rstr import string_repr
- vchunk = inputconst(string_repr, thing)
+ from pypy.rpython.lltypesystem.rstr import string_repr, unicode_repr
+ if is_unicode:
+ vchunk = inputconst(unicode_repr, thing)
+ else:
+ vchunk = inputconst(string_repr, thing)
i = inputconst(Signed, i)
+ if is_unicode and vchunk.concretetype != Ptr(UNICODE):
+ # if we are here, one of the ll_str.* functions returned some
+ # STR, so we convert it to unicode. It's a bit suboptimal
+ # because we do one extra copy.
+ vchunk = hop.gendirectcall(cls.ll_str2unicode, vchunk)
hop.genop('setarrayitem', [vtemp, i, vchunk])
hop.exception_cannot_occur() # to ignore the ZeroDivisionError of '%'
@@ -1009,6 +1036,7 @@
do_stringformat = classmethod(do_stringformat)
TEMP = GcArray(Ptr(STR))
+TEMP_UNICODE = GcArray(Ptr(UNICODE))
# ____________________________________________________________
diff --git a/pypy/rpython/ootypesystem/rstr.py b/pypy/rpython/ootypesystem/rstr.py
--- a/pypy/rpython/ootypesystem/rstr.py
+++ b/pypy/rpython/ootypesystem/rstr.py
@@ -1,4 +1,5 @@
from pypy.tool.pairtype import pairtype
+from pypy.annotation import model as annmodel
from pypy.rlib.rarithmetic import ovfcheck
from pypy.rpython.error import TyperError
from pypy.rpython.rstr import AbstractStringRepr,AbstractCharRepr,\
@@ -79,6 +80,12 @@
sb.ll_append_char(cast_primitive(Char, c))
return sb.ll_build()
+ def ll_unicode(self, s):
+ if s:
+ return s
+ else:
+ return self.convert_const(u'None')
+
def ll_encode_latin1(self, value):
sb = ootype.new(ootype.StringBuilder)
length = value.ll_strlen()
@@ -312,6 +319,7 @@
string_repr = hop.rtyper.type_system.rstr.string_repr
s_str = hop.args_s[0]
assert s_str.is_constant()
+ is_unicode = isinstance(s_str, annmodel.SomeUnicodeString)
s = s_str.const
c_append = hop.inputconst(ootype.Void, 'll_append')
@@ -320,8 +328,15 @@
c8 = hop.inputconst(ootype.Signed, 8)
c10 = hop.inputconst(ootype.Signed, 10)
c16 = hop.inputconst(ootype.Signed, 16)
- c_StringBuilder = hop.inputconst(ootype.Void, ootype.StringBuilder)
- v_buf = hop.genop("new", [c_StringBuilder], resulttype=ootype.StringBuilder)
+ if is_unicode:
+ StringBuilder = ootype.UnicodeBuilder
+ RESULT = ootype.Unicode
+ else:
+ StringBuilder = ootype.StringBuilder
+ RESULT = ootype.String
+
+ c_StringBuilder = hop.inputconst(ootype.Void, StringBuilder)
+ v_buf = hop.genop("new", [c_StringBuilder], resulttype=StringBuilder)
things = cls.parse_fmt_string(s)
argsiter = iter(sourcevarsrepr)
@@ -331,7 +346,12 @@
vitem, r_arg = argsiter.next()
if not hasattr(r_arg, 'll_str'):
raise TyperError("ll_str unsupported for: %r" % r_arg)
- if code == 's' or (code == 'r' and isinstance(r_arg, InstanceRepr)):
+ if code == 's':
+ if is_unicode:
+ vchunk = hop.gendirectcall(r_arg.ll_unicode, vitem)
+ else:
+ vchunk = hop.gendirectcall(r_arg.ll_str, vitem)
+ elif code == 'r' and isinstance(r_arg, InstanceRepr):
vchunk = hop.gendirectcall(r_arg.ll_str, vitem)
elif code == 'd':
assert isinstance(r_arg, IntegerRepr)
@@ -348,13 +368,19 @@
else:
raise TyperError, "%%%s is not RPython" % (code, )
else:
- vchunk = hop.inputconst(string_repr, thing)
- #i = inputconst(Signed, i)
- #hop.genop('setarrayitem', [vtemp, i, vchunk])
+ if is_unicode:
+ vchunk = hop.inputconst(unicode_repr, thing)
+ else:
+ vchunk = hop.inputconst(string_repr, thing)
+ if is_unicode and vchunk.concretetype != ootype.Unicode:
+ # if we are here, one of the ll_str.* functions returned some
+ # STR, so we convert it to unicode. It's a bit suboptimal
+ # because we do one extra copy.
+ vchunk = hop.gendirectcall(cls.ll_str2unicode, vchunk)
hop.genop('oosend', [c_append, v_buf, vchunk], resulttype=ootype.Void)
hop.exception_cannot_occur() # to ignore the ZeroDivisionError of '%'
- return hop.genop('oosend', [c_build, v_buf], resulttype=ootype.String)
+ return hop.genop('oosend', [c_build, v_buf], resulttype=RESULT)
do_stringformat = classmethod(do_stringformat)
diff --git a/pypy/rpython/rpbc.py b/pypy/rpython/rpbc.py
--- a/pypy/rpython/rpbc.py
+++ b/pypy/rpython/rpbc.py
@@ -11,7 +11,7 @@
mangle, inputdesc, warning, impossible_repr
from pypy.rpython import rclass
from pypy.rpython import robject
-from pypy.rpython.annlowlevel import llstr
+from pypy.rpython.annlowlevel import llstr, llunicode
from pypy.rpython import callparse
diff --git a/pypy/rpython/rstr.py b/pypy/rpython/rstr.py
--- a/pypy/rpython/rstr.py
+++ b/pypy/rpython/rstr.py
@@ -483,6 +483,8 @@
# xxx suboptimal, maybe
return str(unicode(ch))
+ def ll_unicode(self, ch):
+ return unicode(ch)
class __extend__(AbstractCharRepr,
AbstractUniCharRepr):
diff --git a/pypy/rpython/test/test_runicode.py b/pypy/rpython/test/test_runicode.py
--- a/pypy/rpython/test/test_runicode.py
+++ b/pypy/rpython/test/test_runicode.py
@@ -1,3 +1,4 @@
+# -*- encoding: utf-8 -*-
from pypy.rpython.lltypesystem.lltype import malloc
from pypy.rpython.lltypesystem.rstr import LLHelpers, UNICODE
@@ -194,7 +195,16 @@
assert self.interpret(fn, [u'(']) == False
assert self.interpret(fn, [u'\u1058']) == False
assert self.interpret(fn, [u'X']) == True
-
+
+ def test_strformat_unicode_arg(self):
+ const = self.const
+ def percentS(s):
+ return const("before %s after") % (s,)
+ #
+ res = self.interpret(percentS, [const(u'à')])
+ assert self.ll_to_string(res) == const(u'before à after')
+ #
+
def unsupported(self):
py.test.skip("not supported")
@@ -202,12 +212,6 @@
test_upper = unsupported
test_lower = unsupported
test_splitlines = unsupported
- test_strformat = unsupported
- test_strformat_instance = unsupported
- test_strformat_nontuple = unsupported
- test_percentformat_instance = unsupported
- test_percentformat_tuple = unsupported
- test_percentformat_list = unsupported
test_int = unsupported
test_int_valueerror = unsupported
test_float = unsupported
More information about the pypy-commit
mailing list