summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Fix a clang warning in grammar.cVictor Stinner2016-08-191-0/+6
| | | | | Clang is smarter than GCC and emits a warning for dead code after a function declared with __attribute__((__noreturn__)) (Py_FatalError).
* Issue #12946: Remove dead code in PyModule_GetDictBerker Peksag2016-08-191-2/+1
| | | | | | PyModule_NewObject already sets md_dict to PyDict_New(): m->md_dict = PyDict_New();
* Issue #12946: Merge from 3.5Berker Peksag2016-08-191-4/+7
|\
| * Issue #12946: Document that PyModule_GetDict can fail in some casesBerker Peksag2016-08-191-4/+7
| |
* | Issue #27157: Make only type() itself accept the one-argument formBerker Peksag2016-08-194-4/+30
| | | | | | | | Patch by Eryk Sun and Emanuel Barry.
* | Rewrap long lines in Misc/NEWS.R David Murray2016-08-191-30/+38
|\ \ | |/ | | | | Also moved news item for #2466 to the correct place.
| * Rewrap long lines in Misc/NEWS.R David Murray2016-08-191-20/+26
| |
* | Merge: #2466: ismount now recognizes mount points user can't access.R David Murray2016-08-194-0/+27
|\ \ | |/
| * #2466: ismount now recognizes mount points user can't access.R David Murray2016-08-194-2/+27
| | | | | | | | | | Patch by Robin Roth, reviewed by Serhiy Storchaka, comment wording tweaked by me.
* | Updated NEWS with information on the argparse change.Vinay Sajip2016-08-181-0/+3
| |
* | Closes #12713: Allowed abbreviation of subcommands in argparse.Vinay Sajip2016-08-183-32/+95
| |
* | Anti-registration of various ABC methods.Guido van Rossum2016-08-1815-62/+300
| | | | | | | | | | | | | | | | | | | | - Issue #25958: Support "anti-registration" of special methods from various ABCs, like __hash__, __iter__ or __len__. All these (and several more) can be set to None in an implementation class and the behavior will be as if the method is not defined at all. (Previously, this mechanism existed only for __hash__, to make mutable classes unhashable.) Code contributed by Andrew Barnert and Ivan Levkivskyi.
* | Merge 3.5 (fix raise)Victor Stinner2016-08-183-1/+24
|\ \ | |/
| * Fix SystemError in "raise" statementVictor Stinner2016-08-183-1/+23
| | | | | | | | | | | | | | | | Issue #27558: Fix a SystemError in the implementation of "raise" statement. In a brand new thread, raise a RuntimeError since there is no active exception to reraise. Patch written by Xiang Zhang.
* | Issue #16764: Move NEWS entry to correct section and remove too strict test.Serhiy Storchaka2016-08-182-7/+3
| |
* | Issue #24773: Include Tallinn 1999-10-31 transition in tests.Alexander Belopolsky2016-08-171-3/+0
| | | | | | | | | | Does not appear to be a problem anymore and I cannot figure out why it was skipped in the first place.
* | Issue #27594: Prevent assertion error when running test_ast with coverageNed Deily2016-08-172-1/+5
| | | | | | | | | | enabled: ensure code object has a valid first line number. Patch suggested by Ivan Levkivskyi.
* | Issue #27786: Simplify x_sub()Victor Stinner2016-08-171-3/+1
| | | | | | | | | | The z variable is known to be a fresh number which cannot be shared, Py_SIZE() can be used directly to negate the number.
* | Closes #9998: Allowed find_library to search additional locations for libraries.Vinay Sajip2016-08-174-6/+82
| |
* | Merge with 3.5Zachary Ware2016-08-171-2/+2
|\ \ | |/
| * Use sys.version_info, not sys.version.Zachary Ware2016-08-171-2/+2
| | | | | | | | | | sys.version[0] gives a string, which fails > comparison with 2. Reported by Arne Maximilian Richter on docs@
* | regrtest: add a summary of the summary, "Result: xxx"Victor Stinner2016-08-172-4/+25
| | | | | | | | | | | | | | It's sometimes hard to check quickly if tests succeeded, failed or something bad happened. I added a final "Result: xxx" line which summarizes all outputs into a single line, written at the end (it should always be the last line of the output).
* | regrtest: set interrupted to True if re-run is interruptedVictor Stinner2016-08-171-1/+2
| |
* | regrtest: add newlines in output for readabilityVictor Stinner2016-08-171-0/+3
| |
* | Merge 3.5 (socket.__all__)Victor Stinner2016-08-171-0/+1
|\ \ | |/
| * Issue #27698: Add socketpair to socket.__all__ on WindowsVictor Stinner2016-08-171-0/+1
| |
* | Null mergeVictor Stinner2016-08-170-0/+0
|\ \ | |/
| * Issue #27726: Fix "make tags"Victor Stinner2016-08-171-3/+3
| | | | | | | | | | | | | | | | * Memove -t option of ctags. The option was kept for backward compatibility, but it was completly removed recently. Patch written by Stéphane Wirtel. * Set locale to C to call sort. vim expects that the tags file is sorted using english collation, so it fails if the locale is french for example. Use LC_ALL=C to force english sorting order. .
* | Fix "make tags": set locale to C to call sortVictor Stinner2016-08-171-1/+1
| | | | | | | | | | | | vim expects that the tags file is sorted using english collation, so it fails if the locale is french for example. Use LC_ALL=C to force english sorting order. Issue #27726.
* | "make tags": remove -t option of ctagsVictor Stinner2016-08-171-2/+2
| | | | | | | | | | The option was kept for backward compatibility, but it was completly removed recently. Patch written by Stéphane Wirtel.
* | Merge 3.5 (script_helper)Victor Stinner2016-08-171-8/+8
|\ \ | |/
| * script_helper: kill the subprocess on errorVictor Stinner2016-08-171-8/+8
| | | | | | | | | | | | | | | | If Popen.communicate() raises an exception, kill the child process to not leave a running child process in background and maybe create a zombi process. This change fixes a ResourceWarning in Python 3.6 when unit tests are interrupted by CTRL+c.
* | regrtest: nicer output for durationsVictor Stinner2016-08-172-14/+21
| | | | | | | | Use milliseconds and minutes units, not only seconds.
* | Tests: add --slowest option to buildbotsVictor Stinner2016-08-172-2/+2
| | | | | | | | Display the top 10 slowest tests.
* | regrtest: rename --slow option to --slowestVictor Stinner2016-08-172-7/+7
| | | | | | | | Thanks to optparse, --slow syntax still works ;-)
* | Fix typo in test_time.pyVictor Stinner2016-08-171-1/+1
| |
* | Minor readability tweakRaymond Hettinger2016-08-171-1/+1
| |
* | merge 3.5 (#27783)Benjamin Peterson2016-08-172-8/+10
|\ \ | |/
| * merge 3.4 (#27783)Benjamin Peterson2016-08-172-8/+10
| |\
| | * merge 3.3 (#27783)Benjamin Peterson2016-08-172-8/+9
| | |\
| | | * rearrange methodcaller_new so that the main error case does not cause ↵Benjamin Peterson2016-08-172-8/+9
| | | | | | | | | | | | | | | | uninitialized memory usage (closes #27783)
* | | | Use Py_ssize_t in _PyEval_EvalCodeWithName()Victor Stinner2016-08-161-8/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Issue #27128, #18295: replace int type with Py_ssize_t for index variables used for positional arguments. It should help to avoid integer overflow and help to emit better machine code for "i++" (no trap needed for overflow). Make also the total_args variable constant.
* | | | Issue #27128: Cleanup _PyEval_EvalCodeWithName()Victor Stinner2016-08-161-11/+43
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * Add comments * Add empty lines for readability * PEP 7 style for if block * Remove useless assert(globals != NULL); (globals is tested a few lines before)
* | | | Issue #27181: Skip tests known to fail until a fix is foundVictor Stinner2016-08-161-0/+1
| | | |
* | | | Add cross-reference to typing.NamedTuple. Doctest two more examples.Raymond Hettinger2016-08-161-4/+10
| | | |
* | | | Add a missing #Brett Cannon2016-08-161-1/+1
| | | |
* | | | fix hyperlinkEthan Furman2016-08-161-1/+1
| | | |
* | | | mergeRaymond Hettinger2016-08-161-5/+0
|\ \ \ \ | |/ / /
| * | | Remove duplicate of the entry in the seealso sectionRaymond Hettinger2016-08-161-5/+0
| | | |
* | | | Issue #25628: Make namedtuple "rename" and "verbose" parameters keyword-only.Raymond Hettinger2016-08-164-3/+22
| | | |