| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Correct test suite for #848017. | Georg Brandl | 2005-08-25 | 1 | -3/+3 |
| | | |||||
| * | Exercise sorted() where possible | Raymond Hettinger | 2004-01-04 | 1 | -3/+1 |
| | | |||||
| * | Suppress unsafe *Cookie class warnings | Andrew M. Kuchling | 2002-12-29 | 1 | -0/+5 |
| | | |||||
| * | Get rid of relative imports in all unittests. Now anything that | Barry Warsaw | 2002-07-23 | 1 | -1/+1 |
| | | | | | | | | | | | | imports e.g. test_support must do so using an absolute package name such as "import test.test_support" or "from test import test_support". This also updates the README in Lib/test, and gets rid of the duplicate data dirctory in Lib/test/data (replaced by Lib/email/test/data). Now Tim and Jack can have at it. :) | ||||
| * | Teach regrtest how to pass on doctest failure msgs. This is done via a | Tim Peters | 2001-09-09 | 1 | -4/+2 |
| | | | | | | | | | | | | horridly inefficient hack in regrtest's Compare class, but it's about as clean as can be: regrtest has to set up the Compare instance before importing a test module, and by the time the module *is* imported it's too late to change that decision. The good news is that the more tests we convert to unittest and doctest, the less the inefficiency here matters. Even now there are few tests with large expected-output files (the new cost here is a Python-level call per .write() when there's an expected- output file). | ||||
| * | Get rid of the superstitious "~" in dict hashing's "i = (~hash) & mask". | Tim Peters | 2001-05-13 | 1 | -1/+3 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 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 Guido fiddled Cookie.py to work with doctest, it's a Good Thing to | Tim Peters | 2001-04-06 | 1 | -0/+4 |
| | | | | | have the std test suite exercise the Cookie doctests too. | ||||
| * | Added test for patch #103473: test an unquoted cookie value containing '=' | Andrew M. Kuchling | 2001-02-21 | 1 | -0/+3 |
| | | |||||
| * | a bold attempt to fix things broken by MAL's verify patch: import | Fredrik Lundh | 2001-01-17 | 1 | -0/+1 |
| | | | | | 'verify' iff it's used by a test module... | ||||
| * | Marc-Andre must not have run these tests -- they used verify() but | Guido van Rossum | 2001-01-17 | 1 | -0/+1 |
| | | | | | | didn't import it. Also got rid of some inconsistent spaces inside parentheses in test_gzip.py. | ||||
| * | This patch removes all uses of "assert" in the regression test suite | Marc-André Lemburg | 2001-01-17 | 1 | -7/+7 |
| | | | | | | | | and replaces them with a new API verify(). As a result the regression suite will also perform its tests in optimization mode. Written by Marc-Andre Lemburg. Copyright assigned to Guido van Rossum. | ||||
| * | Make reindent.py happy (convert everything to 4-space indents!). | Fred Drake | 2000-10-23 | 1 | -4/+2 |
| | | |||||
| * | Updated test suite: test repr() and str() of cookies, and test metadata | Andrew M. Kuchling | 2000-08-24 | 1 | -1/+10 |
| | | | | | fields with quoted values (as in Path="/acme") | ||||
| * | Adding tests of the "attrs" optional argument, and of the js_output | Moshe Zadka | 2000-08-19 | 1 | -0/+5 |
| | | | | | functionality. | ||||
| * | Test case for Cookie.py | Andrew M. Kuchling | 2000-08-19 | 1 | -0/+26 |
 |
index : cpython.git | |
| https://github.com/python/cpython.git |
| summaryrefslogtreecommitdiffstats |