homepage

There are at least 4 places in the standard Python 2.3 library which 
build an RFC 2822 formatted string:
rfc822.formatstring
email.Utils.formatdate
logging.handlers.SMTPHandler.date_time
BaseHTTPServer.HTTPServer.date_time_string
Looking at them, it makes sense to me to
 - replace rfc822's implementation with email's
 (that's the most flexible of the bunch)
 - start a migration so that email uses the one
 from rfc822, if available, else it's own implementation
 (since email is distributed to older Pythons)
 - have logging use the one in rfc822.
 - have BaseHTTPServer use the one in rfc822
If this is deemed an appropriate change, I can send in
a patch.

Logged In: YES 
user_id=29957
It's not just formatdate - various email formatting tasks
are probably the
single largest source of duplication in the std library...

Logged In: YES 
user_id=12800
Personally, I'd like to see the email package's date formatting (and other 
email related tasks) be the canonical standard, and for other modules to use 
its rules when appropriate.

Logged In: YES 
user_id=139865
Careful, date/time formats differ between RFCs.
RFC 850 = Wdy, DD-Mon-YYYY HH:MM:SS (UT/GMT/(+/-)DDDD/...)
RFC 2616 = Wdy, DD Mon YYYY HH:MM:SS GMT
RFC 2822 = Wdy, DD Mon YYYY HH:MM:SS (+/-)DDDD
rfc822.formatstring
 deprecated; let it rot
email.Utils.formatdate
 uses RFC 2822, the One True Date Format
logging.handlers.SMTPHandler.date_time
 uses RFC 2616, should be using RFC 2822
BaseHTTPServer.HTTPServer.date_time_string
 uses RFC 2616
Cookie._getdate
 uses RFC 850
Patch 791776 changes SMTPHandler to use rfc2822.formatdate.
 grepping for day names, month names, year, gmtime, and
localtime does not reveal further duplication.

Still present in source, but doesn't seem too bad.

Does this need to be addressed or can it be closed as out of date or what?

It still needs to be addressed. I'm marking it for 3.2 but I doubt it will get addressed before 3.3 in reality. I also made the type 'performance' since we have no 'refactoring' type.

I notice that logging has been updated. Only http.server's date_time_string remains (email.utils.formatdate does support the GMT form).

http://tools.ietf.org/html/draft-ietf-httpbis-p2-semantics-22#section-7.1.1
quoting from HTTP 1.1 bis
 Prior to 1995, there were three different formats commonly used by
 servers to communicate timestamps. For compatibility with old
 implementations, all three are defined here. The preferred format is
 a fixed-length and single-zone subset of the date and time
 specification used by the Internet Message Format [RFC5322].
 HTTP-date = IMF-fixdate / obs-date
 An example of the preferred format is
 1994年11月06日 08:49:37 GMT ; IMF-fixdate
 Examples of the two obsolete formats are
 Sunday, 06-Nov-94 08:49:37 GMT ; obsolete RFC 850 format
 Sun Nov 6 08:49:37 1994 ; ANSI C's asctime() format
 A recipient that parses a timestamp value in an HTTP header field
 MUST accept all three formats. A sender MUST generate the IMF-
 fixdate format when sending an HTTP-date value in a header field.
What http.server.BaseHTTPRequestHandler.date_time_string is currently doing
>>> import time
>>> timestamp = time.time()
>>> weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
>>> monthname = [None,'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun','Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
>>> year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)
>>> s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (weekdayname[wd],day, monthname[month], year,hh, mm, ss)
>>> s
'Mon, 25 Feb 2013 19:26:34 GMT'
what email.utils.formatdate is doing:
>>> import email.utils
>>> email.utils.formatdate(timeval=None,localtime=False, usegmt=True)
'Mon, 25 Feb 2013 19:40:04 GMT'
>>> import time
>>> ts = time.time()
>>> email.utils.formatdate(timeval=ts,localtime=False, usegmt=True)
'Mon, 25 Feb 2013 19:51:50 GMT'
I createad a patch 
s = email.utils.formatdate(timestamp, False, True)
I didn't touch the log method which has a different format which is anyway not compatible with email.utils.

Made a mistake in the previous server.patch, use server.2.patch

Could you regenerate your patch using hg diff? (or at least diff -u)?

R. David Murray:
Sure. Is it better? issue-747320-1.patch
It seems there are already tests for gmt date format in Lib/test/test_email/test_utils.py 

Ok after comments and review by Eric Araujo on the previous patch.
See issue-747320-3.patch
I have ran
→ ./python.exe Lib/test/test_httpservers.py
[...]
----------------------------------------------------------------------
Ran 39 tests in 3.734s
OK
[137158 refs]
That said there is no specific tests for date_time_string() and log_date_time_string(). That might be something to consider.
→ grep date_time Lib/test/test_httpservers.py
[nil]

Tests would be great, especially given that we can add them in 3.2 and when merging 3.2 into 3.3 and then default, it ensures that the new code has no regression.
(A minor thing: I would use "attribute" instead of "variable" in the docstrings.)
There are also test helpers to assert that a warning is sent, but we don’t have a full coverage policy so it’s okay if you don’t feel like adding them.

Eric,
what do you recommend to move forward with this bug and patches?
Need guidance.
Do you have an example for "(A minor thing: I would use "attribute" instead of "variable" in the docstrings.)"
Also which code base I should use? A lot of water has gone under the bridge in one year. :)

> what do you recommend to move forward with this bug and patches?
I would add tests to 3.4, to be sure that changing the code in 3.5 does not break compatibility.
> Do you have an example for "(A minor thing: I would use "attribute" instead of "variable" in the docstrings.)"
"Deprecated weekdayname variable" → "Deprecated weekdayname attribute"

Here is an updated patch (including tests and documentation updates).

- now = time.time()
- year, month, day, hh, mm, ss, x, y, z = time.localtime(now)
- s = "%02d/%3s/%04d %02d:%02d:%02d" % (
- day, self.monthname[month], year, hh, mm, ss)
+ s = time.strftime("%d/%b/%Y %H:%M:%S", time.localtime())
This part of the patch is incorrect. time.strftime() will return non-English values for different locales.
About deprecations of weekdayname and monthname attributes:
I know that they are undocumented, but they are already being used in the wild. For example, see https://github.com/natemago/srv/blob/master/srv.py#L419 Since we don't have any public API for this use case, I'd be +1 to keep them for now.
Here is an updated patch.

New changeset ee64faffd46a by Berker Peksag in branch 'default':
Issue #747320: Use email.utils.formatdate() to avoid code duplication
https://hg.python.org/cpython/rev/ee64faffd46a