[PATCH 2 of 9] test-trusted.py: fix on PyPy

Dan Villiom Podlaski Christiansen danchr at gmail.com
Sun Dec 26 14:13:20 CST 2010


On 26 Dec 2010, at 20:56, Augie Fackler wrote:

> On Dec 26, 2010 2:55 PM, "Dan Villiom Podlaski Christiansen" <
> danchr at gmail.com> wrote:
>>
>> On 26 Dec 2010, at 16:39, Augie Fackler wrote:
>>
>>> On Dec 26, 2010, at 6:51 AM, Dan Villiom Podlaski Christiansen  
>>> wrote:
>>>>
>>>>
>>>> # HG changeset patch
>>>> # User Dan Villiom Podlaski Christiansen <danchr at gmail.com>
>>>> # Date 1293200527 -3600
>>>> # Node ID a37295adc5d962fd8ea86b4dc2472b0ad0f11deb
>>>> # Parent  4233e0fb74d1d801ad2ce9e3f713a0ca4c8ba668
>>>> test-trusted.py: fix on PyPy
>>>>
>>>> The test previously did a few print statements without end-of-line
>>>> followed by a write from the UI object. For some reason, PyPy
>>>> consistently inserted a space one character later. Changing the  
>>>> print
>>>> statements to include a newline fixes this.
>>>
>>>
>>> This sounds like a bug in pypy to me. Do the pypy guys not agree?
>>
>>
>> I just checked, and the warning is printed on standard error,  
>> whereas the
>> dots were printed on standard output. The exact order of intermingled
>> stdout/stderr output seems thoroughly into implementation-defined  
>> territory…
>
> Fair enough.

I asked on #pypy — just in case — and it turns out that they do indeed  
have a bug in this area. This snippet results in different output  
between CPython and PyPy:

import tempfile

s = tempfile.TemporaryFile()

print >> s, '.',
s.write(',')
print >> s, '.',

s.seek(0)
print s.read()

--

Dan Villiom Podlaski Christiansen
danchr at gmail.com

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1943 bytes
Desc: not available
URL: <http://selenic.com/pipermail/mercurial-devel/attachments/20101226/b81f3938/attachment.bin>


More information about the Mercurial-devel mailing list