| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Patch #444359: Remove unused imports. | Martin v. Löwis | 2001-08-02 | 1 | -1/+1 |
| | | |||||
| * | Use string.ascii_letters instead of string.letters (SF bug #226706). | Fred Drake | 2001-07-20 | 1 | -1/+1 |
| | | |||||
| * | Get rid of the superstitious "~" in dict hashing's "i = (~hash) & mask". | Tim Peters | 2001-05-13 | 1 | -6/+14 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 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. | ||||
| * | Since this module already uses doctest-style examples, I figured I'd | Guido van Rossum | 2001-04-06 | 1 | -4/+10 |
| | | | | | | | | | | | | | | | | 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! | ||||
| * | Patch #103473 from dougfort: Some sites (amazon.com for one) drop | Andrew M. Kuchling | 2001-02-20 | 1 | -2/+2 |
| | | | | | | | 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 '='. | ||||
| * | added __all__ lists to a number of Python modules | Skip Montanaro | 2001-01-20 | 1 | -0/+2 |
| | | | | | | | | | 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. | ||||
| * | Whitespace normalization. | Tim Peters | 2001-01-14 | 1 | -14/+14 |
| | | |||||
| * | Update the code to better reflect recommended style: | Fred Drake | 2000-12-12 | 1 | -1/+1 |
| | | | | | | Use != instead of <> since <> is documented as "obsolescent". Use "is" and "is not" when comparing with None or type objects. | ||||
| * | Whitespace cleanup; now passes the regression test (the last checkin made | Fred Drake | 2000-08-24 | 1 | -21/+15 |
| | | | | | | | | 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. | ||||
| * | Updated version of Cookie.py (rev. 2.29) from timo | Andrew M. Kuchling | 2000-08-24 | 1 | -19/+33 |
| | | |||||
| * | Added Tim O'Malley's Cookie.py module (master version at | Andrew M. Kuchling | 2000-08-19 | 1 | -0/+718 |
| http://www.timo-tasi.org/python/Cookie.py) This is revision 2.26 according to Tim's RCS history. | |||||
 |
index : cpython.git | |
| https://github.com/python/cpython.git |
| summaryrefslogtreecommitdiffstats |