Commit Graph

20 Commits

Author SHA1 Message Date
Georg Brandl 8246c439a8 Correct test suite for #848017. 2005-08-25 07:32:42 +00:00
Georg Brandl 532efabf1d patch #848017: make Cookie more RFC-compliant. 2005-08-24 22:34:21 +00:00
Georg Brandl 03a33ea3a8 bug [ 1108948 ] Cookie.py produces invalid code 2005-06-26 21:02:49 +00:00
Walter Dörwald f0dfc7ac5c Fix a bunch of typos in documentation, docstrings and comments.
(From SF patch #810751)
2003-10-20 14:01:56 +00:00
Andrew M. Kuchling 7877a76107 Patch #655760: add warnings when the unsafe *Cookie classes are instantiated 2002-12-29 16:44:31 +00:00
Andrew M. Kuchling 6cd77126b4 No point in warning about needing re module; remove helpful message 2002-12-17 18:59:51 +00:00
Andrew M. Kuchling 3c76ad0f8d Fix comment typo 2002-12-17 18:56:26 +00:00
Raymond Hettinger 0a2963c797 Apply SF 562987 modernizing Cookie to subclass from dict instead of UserDict 2002-06-26 15:19:01 +00:00
Fred Drake d451ec1cdb Clean up uses of some deprecated features.
Reported by Neal Norwitz on python-dev.
2002-04-26 02:29:55 +00:00
Martin v. Löwis 02d893cfae Patch #444359: Remove unused imports. 2001-08-02 07:15:29 +00:00
Fred Drake 79e75e1916 Use string.ascii_letters instead of string.letters (SF bug #226706). 2001-07-20 19:05:50 +00:00
Tim Peters 2f228e75e4 Get rid of the superstitious "~" in dict hashing's "i = (~hash) & mask".
The comment following used to say:
	/* We use ~hash instead of hash, as degenerate hash functions, such
	   as for ints <sigh>, can have lots of leading zeros. It's not
	   really a performance risk, but better safe than sorry.
	   12-Dec-00 tim:  so ~hash produces lots of leading ones instead --
	   what's the gain? */
That is, there was never a good reason for doing it.  And to the contrary,
as explained on Python-Dev last December, it tended to make the *sum*
(i + incr) & mask (which is the first table index examined in case of
collison) the same "too often" across distinct hashes.

Changing to the simpler "i = hash & mask" reduced the number of string-dict
collisions (== # number of times we go around the lookup for-loop) from about
6 million to 5 million during a full run of the test suite (these are
approximate because the test suite does some random stuff from run to run).
The number of collisions in non-string dicts also decreased, but not as
dramatically.

Note that this may, for a given dict, change the order (wrt previous
releases) of entries exposed by .keys(), .values() and .items().  A number
of std tests suffered bogus failures as a result.  For dicts keyed by
small ints, or (less so) by characters, the order is much more likely to be
in increasing order of key now; e.g.,

>>> d = {}
>>> for i in range(10):
...    d[i] = i
...
>>> d
{0: 0, 1: 1, 2: 2, 3: 3, 4: 4, 5: 5, 6: 6, 7: 7, 8: 8, 9: 9}
>>>

Unfortunately. people may latch on to that in small examples and draw a
bogus conclusion.

test_support.py
    Moved test_extcall's sortdict() into test_support, made it stronger,
    and imported sortdict into other std tests that needed it.
test_unicode.py
    Excluced cp875 from the "roundtrip over range(128)" test, because
    cp875 doesn't have a well-defined inverse for unicode("?", "cp875").
    See Python-Dev for excruciating details.
Cookie.py
    Chaged various output functions to sort dicts before building
    strings from them.
test_extcall
    Fiddled the expected-result file.  This remains sensitive to native
    dict ordering, because, e.g., if there are multiple errors in a
    keyword-arg dict (and test_extcall sets up many cases like that), the
    specific error Python complains about first depends on native dict
    ordering.
2001-05-13 00:19:31 +00:00
Guido van Rossum 58b6f5b53e Since this module already uses doctest-style examples, I figured I'd
add a self-test using doctest.  Results:

- The docstring needs to be a raw string because it uses \"...\".

- The oreo example was broken: the Set-Cookie output doesn't add
  quotes around "doublestuff".

- I had to change the example that prints the class of a Cookie.Cookie
  instance to avoid incorporating an arbitrary object address in the
  test output.

Pretty good score for both doctest and the doc string, I'd say!
2001-04-06 19:39:11 +00:00
Andrew M. Kuchling c05abb3bda Patch #103473 from dougfort: Some sites (amazon.com for one) drop
cookies that contain '=' as part of the value. This patch modifies
Cookie.py to allow '=' as a legal character, and to make the key
search nongreedy so it stops at the first '='.
2001-02-20 22:11:24 +00:00
Skip Montanaro e99d5ea25b added __all__ lists to a number of Python modules
added test script and expected output file as well
this closes patch 103297.
__all__ attributes will be added to other modules without first submitting
a patch, just adding the necessary line to the test script to verify
more-or-less correct implementation.
2001-01-20 19:54:20 +00:00
Tim Peters 88869f9787 Whitespace normalization. 2001-01-14 23:36:06 +00:00
Fred Drake 8152d32375 Update the code to better reflect recommended style:
Use != instead of <> since <> is documented as "obsolescent".
Use "is" and "is not" when comparing with None or type objects.
2000-12-12 23:20:45 +00:00
Fred Drake ff5364ac9d Whitespace cleanup; now passes the regression test (the last checkin made
it fail on a TabError (inconsistent tab/space usage)).

Removed a comment about including a test since there is a regression test
for this module.
2000-08-24 14:40:35 +00:00
Andrew M. Kuchling 0b29b11187 Updated version of Cookie.py (rev. 2.29) from timo 2000-08-24 11:52:33 +00:00
Andrew M. Kuchling 52ea872777 Added Tim O'Malley's Cookie.py module (master version at
http://www.timo-tasi.org/python/Cookie.py)
This is revision 2.26 according to Tim's RCS history.
2000-08-19 13:01:19 +00:00