Encoding problem when launching Python27 via DOS
I created a simple program which writes in a unicode files some french text
with accents!
*# -*- coding: cp1252 -*-*
*#!/usr/bin/python*
*'''*
*Created on 27 déc. 2010*
*
*
*@author: jpmena*
*'''*
*from datetime import datetime*
*import locale*
*import codecs*
*import os,sys*
*
*
*class Log(object):*
*log=None*
*def __init__(self,log_path,charset_log=None):*
*self.log_path=log_path*
*if(os.path.exists(self.log_path)):*
*os.remove(self.log_path)*
*#self.log=open(self.log_path,'a')*
*if charset_log is None:*
*self.charset_log=sys.getdefaultencoding()*
*else:*
*self.charset_log=charset_log*
*self.log=codecs.open(self.log_path, "a", charset_log)*
**
*def getInstance(log_path=None):*
*print "encodage systeme:"+sys.getdefaultencoding()*
*if Log.log is None:*
*if log_path is None:*
*log_path=os.path.join(os.getcwd(),'logParDefaut.log')*
*Log.log=Log(log_path)*
*return Log.log*
**
*getInstance=staticmethod(getInstance)*
**
*def p(self,msg):*
*aujour_dhui=datetime.now()*
*date_stamp=aujour_dhui.strftime("%d/%m/%y-%H:%M:%S")*
*print sys.getdefaultencoding()*
*unicode_str=u'%s : %s \n' %
(date_stamp,msg.encode(self.charset_log,'replace'))*
*self.log.write(unicode_str)*
*return unicode_str*
**
*def close(self):*
*self.log.flush()*
*self.log.close()*
*return self.log_path*
*
*
*if __name__ == '__main__':*
*l=Log.getInstance()*
*l.p("premier message de Log à accents")*
*Log.getInstance().p("second message de Log")*
*l.close()*
I am using PyDev/Aptana for developping. Il Aptana lanches the program
everything goes well!!! sys.getdefaultencoding() answers 'cp1252'
But if I execute the following batch file in a DOS console on my Windows
VISTA:
*@echo off*
*setlocal*
*chcp 1252*
*set PYTHON_HOME=C:\Python27*
*for /F "tokens=1-4 delims=/ " %%i in ('date /t') do (*
*if "%%l"=="" (*
*:: Windows XP*
*set D=%%k%%j%%i*
* ) else (*
*:: Windows NT/2000*
*set D=%%l%%k%%j*
* )*
*)*
*set PYTHONIOENCODING=cp1252:backslashreplace*
*%PYTHON_HOME%\python.exe "%~dp0\src\utils\Log.py"*
the answer is:
*C:\Users\jpmena\Documents\My
Dropbox\RIF\Python\VelocityTransforms>generationPro*
*grammeSitePublicActuel.cmd*
*Page de codes active : 1252*
*encodage systeme:ascii*
*ascii*
*Traceback (most recent call last):*
* File "C:\Users\jpmena\Documents\My
Dropbox\RIF\Python\VelocityTransforms\\src\*
*utils\Log.py", line 51, in *
*l.p("premier message de Log à accents")*
* File "C:\Users\jpmena\Documents\My
Dropbox\RIF\Python\VelocityTransforms\\src\*
*utils\Log.py", line 40, in p*
*unicode_str=u'%s : %s \n' %
(date_stamp,msg.encode(self.charset_log,'replac*
*e'))*
*UnicodeDecodeError: 'ascii' codec can't decode byte 0xe0 in position 23:
ordinal*
* not in range(128)*
sys.getdefaultencoding answers ascii so the encode function cannot encode
the accent in 'à'
I am using Python27 because it is compatible with the actual versions of
pyodbc (for accessinf a ACCESS database) and airspeed (Velocity Templates in
utf-8)
The target is to launch airspeed applications via the Windows CRON
Can someone help me. I am really stuck!
Thanks...
--
http://mail.python.org/mailman/listinfo/python-list
Re: Encoding problem when launching Python27 via DOS
Thanks a lot for this quick answer! It is very clear!
Ti better understand what the difference between encoding and decoding is I
found the following website: http://www.evanjones.ca/python-utf8.html
<http://www.evanjones.ca/python-utf8.html>I change the program to (changes
are in bold):
*# -*- coding: utf8 -*- *(no more cp1252 the source file is directly in
unicode)
*#!/usr/bin/python*
*'''*
*Created on 27 déc. 2010*
*
*
*@author: jpmena*
*'''*
*from datetime import datetime*
*import locale*
*import codecs*
*import os,sys*
*
*
*class Log(object):*
*log=None*
*def __init__(self,log_path):*
*self.log_path=log_path*
*if(os.path.exists(self.log_path)):*
*os.remove(self.log_path)*
*#self.log=open(self.log_path,'a')*
*self.log=codecs.open(self.log_path, "a", 'utf-8')*
**
*def getInstance(log_path=None):*
*print "encodage systeme:"+sys.getdefaultencoding()*
*if Log.log is None:*
*if log_path is None:*
*log_path=os.path.join(os.getcwd(),'logParDefaut.log')*
*Log.log=Log(log_path)*
*return Log.log*
**
*getInstance=staticmethod(getInstance)*
**
*def p(self,msg):*
*aujour_dhui=datetime.now()*
*date_stamp=aujour_dhui.strftime("%d/%m/%y-%H:%M:%S")*
*print sys.getdefaultencoding()*
*unicode_str='%s : %s \n' % (date_stamp,unicode(msg,'utf-8'))*
*#unicode_str=msg*
*self.log.write(unicode_str)*
*return unicode_str*
**
*def close(self):*
*self.log.flush()*
*self.log.close()*
*return self.log_path*
*
*
*if __name__ == '__main__':*
*l=Log.getInstance()*
*l.p("premier message de Log à accents")*
*Log.getInstance().p("second message de Log")*
*l.close()*
The DOS conole output is now:
*C:\Documents and Settings\jpmena\Mes
documents\VelocityRIF\VelocityTransforms>generationProgrammeSitePublicActuel.cmd
*
*Page de codes active : 1252*
*encodage systeme:ascii*
*ascii*
*encodage systeme:ascii*
*ascii*
And the Generated Log file showsnow the expected result:
*11/04/11-10:53:44 : premier message de Log à accents *
*11/04/11-10:53:44 : second message de Log*
Thanks.
If you have other links of interests about unicode encoding and decoding in
Python. They are welcome
2011/4/10 MRAB
> On 10/04/2011 13:22, Jean-Pierre M wrote:
> > I created a simple program which writes in a unicode files some french
> text with accents!
> [snip]
> This line:
>
>
>l.p("premier message de Log à accents")
>
> passes a bytestring to the method, and inside the method, this line:
>
>
>unicode_str=u'%s : %s \n' %
> (date_stamp,msg.encode(self.charset_log,'replace'))
>
> it tries to encode the bytestring to Unicode.
>
> It's not possible to encode a bytestring, only a Unicode string, so
> Python tries to decode the bytestring using the fallback encoding
> (ASCII) and then encode the result.
>
> Unfortunately, the bytestring isn't ASCII (it contains accented
> characters), so it can't be decoded as ASCII, hence the exception.
>
> BTW, it's probably better to forget about cp1252, etc, and use UTF-8
> instead, and also to use Unicode wherever possible.
> --
> http://mail.python.org/mailman/listinfo/python-list
>
--
http://mail.python.org/mailman/listinfo/python-list
Writing Exif File
I want to enter Comments of a picture in a JPeg file. Is there a library in Python which allow me to do that without having to reinvent the wheel? The target is to read those comments in my private webSite using the php exif_read_data function (http://php.net/manual/fr/book.exif.php) -- http://mail.python.org/mailman/listinfo/python-list
