Author: Edd Barrett <vex...@gmail.com>
Branch:
Changeset: r77837:131622d03fbe
Date: 2015-06-03 17:54 +0100
http://bitbucket.org/pypy/pypy/changeset/131622d03fbe/
Log: Backed out changeset 5cf9f578ca18
This broke translation on our debian 8 machine.
I have tried setting various different LC_ALL/LC_CTYPE values, but
these seem to have no effect.
17:01 < fijal> ebarrett: well then rollback with an explanation
17:01 < fijal> ebarrett: leaving build broken is not a good idea
Translation error follows:
[translation:info] File "/opt/pypy/rpython/annotator/unaryop.py",
line 604, in method_decode [translation:info] raise
AnnotatorError("Encoding %s not supported for strings" % (enc,))
[translation:ERROR] AnnotatorError: [translation:ERROR]
[translation:ERROR] Encoding latin1 not supported for strings
[translation:ERROR] [translation:ERROR] [translation:ERROR] Occurred
processing the following simple_call: [translation:ERROR]
(AttributeError getting at the binding!) [translation:ERROR]
v334 = simple_call(v333, ('latin1')) [translation:ERROR]
[translation:ERROR] In <FunctionGraph of
(pypy.objspace.std.newformat:529)Formatter._lit at 0x3724fa98>:
[translation:ERROR] Happened at file
/opt/pypy/pypy/objspace/std/newformat.py line 531
[translation:ERROR] [translation:ERROR] ==> return
s.decode("latin1") [translation:ERROR] [translation:ERROR] Known
variable annotations: [translation:ERROR] v333 =
SomeBuiltinMethod(analyser=<rpython.tool.descriptor.InstanceMethod
object at 0x000000003802d910>, methodname='decode',
s_self=SomeChar(const=' ', no_nul=True)) [translation:ERROR]
[translation:ERROR] Processing block: [translation:ERROR] block@12
is a <class 'rpython.flowspace.flowcontext.SpamBlock'>
[translation:ERROR] in
(pypy.objspace.std.newformat:529)Formatter._lit [translation:ERROR]
containing the following operations: [translation:ERROR] v333
= getattr(s_11, ('decode')) [translation:ERROR] v334 =
simple_call(v333, ('latin1')) [translation:ERROR] --end--
diff --git a/pypy/objspace/std/newformat.py b/pypy/objspace/std/newformat.py
--- a/pypy/objspace/std/newformat.py
+++ b/pypy/objspace/std/newformat.py
@@ -528,7 +528,7 @@
def _lit(self, s):
if self.is_unicode:
- return s.decode("latin1")
+ return s.decode("ascii")
else:
return s
@@ -586,8 +586,8 @@
thousands = ""
grouping = "\xFF" # special value to mean 'stop'
if self.is_unicode:
- self._loc_dec = dec.decode("latin1")
- self._loc_thousands = thousands.decode("latin1")
+ self._loc_dec = dec.decode("ascii")
+ self._loc_thousands = thousands.decode("ascii")
else:
self._loc_dec = dec
self._loc_thousands = thousands
@@ -725,7 +725,7 @@
out.append_multiple_char(fill_char[0], spec.n_lpadding)
if spec.n_sign:
if self.is_unicode:
- sign = spec.sign.decode("latin1")
+ sign = spec.sign.decode("ascii")
else:
sign = spec.sign
out.append(sign)
@@ -828,14 +828,14 @@
prefix = "0x"
as_str = value.format(LONG_DIGITS[:base], prefix)
if self.is_unicode:
- return as_str.decode("latin1")
+ return as_str.decode("ascii")
return as_str
def _int_to_base(self, base, value):
if base == 10:
s = str(value)
if self.is_unicode:
- return s.decode("latin1")
+ return s.decode("ascii")
return s
# This part is slow.
negative = value < 0
@@ -954,7 +954,7 @@
have_dec_point, to_remainder = self._parse_number(result,
to_number)
n_remainder = len(result) - to_remainder
if self.is_unicode:
- digits = result.decode("latin1")
+ digits = result.decode("ascii")
else:
digits = result
spec = self._calc_num_width(0, sign, to_number, n_digits,
@@ -1059,8 +1059,8 @@
to_imag_number)
if self.is_unicode:
- re_num = re_num.decode("latin1")
- im_num = im_num.decode("latin1")
+ re_num = re_num.decode("ascii")
+ im_num = im_num.decode("ascii")
#set remainder, in CPython _parse_number sets this
#using n_re_digits causes tests to fail
_______________________________________________
pypy-commit mailing list
pypy-commit@python.org
https://mail.python.org/mailman/listinfo/pypy-commit
