| msg109621 - (view) |
Author: norbidur (norbidur) |
Date: 2010-07-08 22:16 |
| SMTPHandler fails when receiving unicode strings. example : import logging,logging.handlers smtpHandler = logging.handlers.SMTPHandler( mailhost=("smtp.free.fr",25), fromaddr="from@free.fr", toaddrs="to@free.fr", subject=u"error message") LOG = logging.getLogger() LOG.addHandler(smtpHandler) LOG.error(u"accentu\u00E9") -> UnicodeEncodeError: 'ascii' codec can't encode character u'\xe9' in position 108 : ordinal not in range(128) There has been a discuss on this in http://groups.google.com/group/comp.lang.python/browse_thread/thread/759df42f9374d1b6/05ad55c388c746e3?lnk=raot&pli=1 FileHandler does not behave the same way : for this handler's family an encoding can be specified, and if this encoding fails, there is a fallback to UTF-8. |
|
|
| msg114691 - (view) |
Author: Vinay Sajip (vinay.sajip) *  |
Date: 2010-08-22 18:34 |
| SMTPHandler provides an implementation for the simplest/most common case. Full support for encoding in emails is likely to be application-specific, i.e. no one-size-fits-all can be easily specified. For example, different encodings could be used for headers, subject and body - say, quoted-printable for the body and base64 for the subject. Unfortunately, support for quoted-printable requires a global state change, see for instance http://radix.twistedmatrix.com/2010/07/how-to-send-good-unicode-email-with.html so it seems not to be a good idea to implement in the logging package itself. I would suggest implementing a handler which subclasses SMTPHandler and does the appropriate formatting as per (for example) the above post. To facilitate this, I coukd add two methods to SMTPHandler: class SMTPHandler(logging.Handler): def prepareEmail(self, record): """ Prepare a record for emailing, including setting up mail headers doing all necessary encodings. Return a value suitable for passing to the sendEmail method. The default implementation will assume all-ASCII content for headers and body, do no special encoding, and return a string. """ def sendMail(self, smtp, msg): """ Send a message via the provided SMTP instance. The default implementation will call smtplib.sendmail(), passing the result from the prepareEmail method. """ I'm not yet sure if this would meet all use cases. Marking as pending, awaiting feedback. Will close in a couple of weeks if no feedback received. |
|
|
| msg114924 - (view) |
Author: R. David Murray (r.david.murray) * |
Date: 2010-08-25 18:11 |
| Given Vinay's last comment I don't think this needs addressing in 2.x, and it is not a problem in 3.x. |
|
|
| msg151524 - (view) |
Author: Chris Withers (cjw296) * |
Date: 2012-01-18 08:45 |
| Just as a post-fix to this, the email handlers for the python logging framework that I maintain as a package on PyPI now handle unicode email correctly: http://pypi.python.org/pypi/mailinglogger/3.7.0 I'd suggest people looking for fully-featured email log handlers use mailinglogger... |
|
|
| msg252928 - (view) |
Author: Simon04 (simon04) * |
Date: 2015-10-13 13:09 |
| I don't see why/how this should be fixed in Python 3. Using the example from and Python 3.5.0, I get: --- Logging error --- Traceback (most recent call last): File "/usr/lib/python3.5/logging/handlers.py", line 985, in emit smtp.sendmail(self.fromaddr, self.toaddrs, msg) File "/usr/lib/python3.5/smtplib.py", line 846, in sendmail msg = _fix_eols(msg).encode('ascii') UnicodeEncodeError: 'ascii' codec can't encode character '\xe9' in position 108: ordinal not in range(128) Call stack: File "/tmp/x.py", line 8, in LOG.error(u"accentu\u00E9") Message: 'accentué' Arguments: () The problem is that an SMTP message is constructed and non-ASCII characters are not escaped in SMTPHandler.emit. A robust fix would be to use email.mime.text.MIMEText instead: msg = MIMEText(msg) msg['Subject'] = self.getSubject(record) msg['From'] = self.fromaddr msg['To'] = ",".join(self.toaddrs) |
|
|
| msg252931 - (view) |
Author: R. David Murray (r.david.murray) * |
Date: 2015-10-13 13:36 |
| In 3.4/3.5 a better fix would be to use EmailMessage instead of Message, and smtp.send_message instead of smtp.sendmail. That will do the right thing, where "the right thing" is defined as defaulting to utf-8 for both headers and body. A specific application might want other defaults, in which case one should proceed as Vinay suggests. You could open a new issue requesting the use of EmailMessage/send_message in python3 logging if you want to pursue this. Note that this would also allow unicode in the display name part of the addresses. |
|
|
| msg252932 - (view) |
Author: R. David Murray (r.david.murray) * |
Date: 2015-10-13 13:38 |
| To clarify: it will default to utf-8 encoded for email transport. (Since email now supports SMTPUTF8, what I said could have meant defaulting to that, which it does *not* do.) |
|
|