0287003 | Jeff Whitaker | 29 November 2016, 17:34:07 UTC | Merge pull request #607 from Unidata/v1.2.5rel version 1.2.5 release | 29 November 2016, 17:34:07 UTC |
b9739bf | Jeff Whitaker | 29 November 2016, 04:39:49 UTC | update html docs | 29 November 2016, 04:39:49 UTC |
7164e6a | Jeff Whitaker | 28 November 2016, 20:29:23 UTC | use different dap test server (a more reliable one) | 28 November 2016, 20:29:23 UTC |
bf2868b | Jeff Whitaker | 28 November 2016, 18:28:15 UTC | fix so that tutorial.py works | 28 November 2016, 18:28:15 UTC |
9e05ede | Jeff Whitaker | 28 November 2016, 16:51:40 UTC | update | 28 November 2016, 16:51:40 UTC |
98fdfb5 | Jeff Whitaker | 28 November 2016, 16:51:22 UTC | update for version 1.2.5 | 28 November 2016, 16:51:22 UTC |
ed06bad | Jeff Whitaker | 28 November 2016, 16:47:26 UTC | regenerate C | 28 November 2016, 16:47:26 UTC |
2e010db | Jeff Whitaker | 28 November 2016, 16:33:04 UTC | release branch for version 1.2.5 | 28 November 2016, 16:33:04 UTC |
1be10e8 | Jeff Whitaker | 28 November 2016, 16:29:29 UTC | Merge pull request #605 from ckhroulev/issue527 Initialize members of vldata to make sure that nc_free_vlens() succeeds. See #527. | 28 November 2016, 16:29:29 UTC |
1db8833 | Constantine Khroulev | 23 November 2016, 23:38:15 UTC | Disable #527 regression test for NetCDF versions older than 4.4.1. | 23 November 2016, 23:38:15 UTC |
b868b3e | Constantine Khroulev | 17 November 2016, 07:17:58 UTC | regenerate C source | 17 November 2016, 07:32:08 UTC |
118477a | Constantine Khroulev | 17 November 2016, 03:31:39 UTC | Update Changelog. | 17 November 2016, 07:32:08 UTC |
7ba7651 | Constantine Khroulev | 17 November 2016, 03:28:35 UTC | Add a regression test for issue #527. | 17 November 2016, 07:18:41 UTC |
ac3d62f | Constantine Khroulev | 16 November 2016, 09:05:13 UTC | Initialize members of vldata to make sure nc_free_vlens() succeeds. Note that free(NULL) is a no-op. | 16 November 2016, 09:05:13 UTC |
d88be75 | Jeff Whitaker | 08 November 2016, 17:19:18 UTC | Merge pull request #604 from Unidata/pickle_error raise NotImplementedError when trying to pickle extension objects (issue #602) | 08 November 2016, 17:19:18 UTC |
d4bd409 | Jeff Whitaker | 07 November 2016, 20:24:46 UTC | regenerate C | 07 November 2016, 20:24:46 UTC |
64770a7 | Jeff Whitaker | 07 November 2016, 20:23:57 UTC | Merge branch 'master' into pickle_error | 07 November 2016, 20:23:57 UTC |
e47e470 | Jeff Whitaker | 07 November 2016, 20:19:31 UTC | Merge pull request #603 from Unidata/negyear fix for neg years in utime.num2date (issue #596) | 07 November 2016, 20:19:31 UTC |
b7aa12f | Jeff Whitaker | 07 November 2016, 20:19:26 UTC | regenerate C source | 07 November 2016, 20:19:26 UTC |
2eada61 | Jeff Whitaker | 07 November 2016, 17:54:56 UTC | move __reduce__ from _Dimension to MFDataset as intended | 07 November 2016, 17:54:56 UTC |
60f8488 | Jeff Whitaker | 07 November 2016, 17:13:52 UTC | update | 07 November 2016, 17:13:52 UTC |
9a4f084 | Jeff Whitaker | 07 November 2016, 17:11:44 UTC | add NotImplementedError when trying to pickle extension objects (issue | 07 November 2016, 17:11:44 UTC |
c1d1dc9 | Jeff Whitaker | 07 November 2016, 16:06:38 UTC | test for utime.num2date with negative year. | 07 November 2016, 16:06:38 UTC |
8cb26ba | Jeff Whitaker | 07 November 2016, 15:49:24 UTC | fix for other problem in issue #596 (negative years fail in utime.num2date) | 07 November 2016, 15:49:24 UTC |
57da8c1 | Jeff Whitaker | 07 November 2016, 15:38:19 UTC | Merge pull request #598 from Unidata/issue596 fix for issue #596 | 07 November 2016, 15:38:19 UTC |
9d68662 | Jeff Whitaker | 04 November 2016, 18:37:36 UTC | remove trailing space | 04 November 2016, 18:37:36 UTC |
f4d8ac9 | Jeff Whitaker | 04 November 2016, 17:54:46 UTC | add test for issue #596 | 04 November 2016, 17:54:46 UTC |
b87fd03 | Jeff Whitaker | 04 November 2016, 17:31:40 UTC | fix for issue #596 | 04 November 2016, 17:31:40 UTC |
0c0c100 | Jeff Whitaker | 03 November 2016, 19:07:34 UTC | Merge pull request #597 from ckhroulev/string-array-attributes Support reading and writing string array attributes. See #592. | 03 November 2016, 19:07:34 UTC |
4086390 | Jeff Whitaker | 31 October 2016, 21:01:39 UTC | Merge pull request #594 from ckhroulev/datetime-arithmetic Implement datetime and timedelta arithmetic. See #593. | 31 October 2016, 21:01:39 UTC |
3fa610d | Constantine Khroulev | 31 October 2016, 20:25:58 UTC | Update docstrings. | 31 October 2016, 20:25:58 UTC |
ee65bb7 | Constantine Khroulev | 31 October 2016, 19:20:02 UTC | Use _strencode() to recover Python 3 compatibility. If would be nice to be able to use b'\x00' instead, but Python 2.6 does not have the "b" string prefix. | 31 October 2016, 19:20:02 UTC |
7515af0 | Constantine Khroulev | 31 October 2016, 18:35:14 UTC | Fix typos in comments and strings. | 31 October 2016, 18:35:14 UTC |
88c28a9 | Constantine Khroulev | 31 October 2016, 17:14:19 UTC | Update the setncattr_string docstring, tests, and the Changelog. | 31 October 2016, 17:47:50 UTC |
6ea23de | Constantine Khroulev | 28 October 2016, 23:22:29 UTC | Re-generate _netCDF4.c with zeros in include/constants.pyx | 28 October 2016, 23:22:29 UTC |
6d287cb | Constantine Khroulev | 28 October 2016, 22:51:08 UTC | Commit Cython-generated C code. | 28 October 2016, 22:51:08 UTC |
6cbebf1 | Constantine Khroulev | 28 October 2016, 22:31:33 UTC | setncattr_string: don't convert value to str. | 28 October 2016, 22:35:49 UTC |
4adc570 | Constantine Khroulev | 28 October 2016, 22:30:00 UTC | Add support for sequences of strings to _set_att(). | 28 October 2016, 22:35:49 UTC |
a46217e | Constantine Khroulev | 28 October 2016, 22:25:17 UTC | Isolate the workaround for #485, remove unused variables. | 28 October 2016, 22:25:17 UTC |
f78a66a | Constantine Khroulev | 17 October 2016, 17:03:14 UTC | Support reading string array attributes. | 28 October 2016, 03:51:20 UTC |
2e3a67d | Constantine Khroulev | 27 October 2016, 23:25:36 UTC | Merge branch 'master' into datetime-arithmetic | 27 October 2016, 23:25:36 UTC |
86b0820 | Constantine Khroulev | 27 October 2016, 23:16:25 UTC | Use inheritance ("subclassing") instead of a function pointer ... to chose the right algorithm for adding timedelta to a datetime. Now we have the following classes - DatetimeGregorian - DatetimeProlepticGregorian - DatetimeJulian - DatetimeNoLeap - DatetimeAllLeap - Datetime360Day - datetime (alias for DatetimeProlepticGregorian, created for backward compatibility). | 27 October 2016, 23:16:25 UTC |
b3f12f8 | Jeff Whitaker | 26 October 2016, 15:42:04 UTC | Merge pull request #595 from jjhelmus/filepath_segfault BUG: Increase buffer used to store filepath | 26 October 2016, 15:42:04 UTC |
0964bd5 | Constantine Khroulev | 25 October 2016, 05:30:14 UTC | Use static typing to speed up datetime comparisons. ... by adding datetime.to_tuple(self), which knows the type of "self". | 25 October 2016, 05:30:14 UTC |
03b51bc | Constantine Khroulev | 25 October 2016, 05:17:49 UTC | Revert "Update datetime.__hash__() to include datetime.calendar." This reverts commit 96b7bc08e25f48a0cdf9cbdb965d359a83a90e99. | 25 October 2016, 05:17:49 UTC |
3b7852b | Constantine Khroulev | 25 October 2016, 02:56:22 UTC | Update Changelog. | 25 October 2016, 03:14:16 UTC |
96b7bc0 | Constantine Khroulev | 25 October 2016, 03:10:42 UTC | Update datetime.__hash__() to include datetime.calendar. As far as I can tell there is no reason why hash(netcdftime.datetime(1,1,1)) should be equal to hash(datetime.datetime(1,1,1)). | 25 October 2016, 03:12:27 UTC |
9a51456 | Constantine Khroulev | 25 October 2016, 00:48:40 UTC | Update the datetime doc. string. Also: fix typos in comments and strings. | 25 October 2016, 00:48:40 UTC |
77b9fd3 | Constantine Khroulev | 24 October 2016, 23:49:17 UTC | Trying to support Cython 0.19... | 24 October 2016, 23:49:17 UTC |
38f28b2 | Constantine Khroulev | 24 October 2016, 23:45:33 UTC | Test datetime.__sub__() with different calendars. | 24 October 2016, 23:45:33 UTC |
0392510 | Jonathan Helmus | 24 October 2016, 23:36:48 UTC | Regenerate _netCDF4.c from Cython source | 24 October 2016, 23:36:48 UTC |
48b86e3 | Jonathan Helmus | 24 October 2016, 23:36:22 UTC | Add Changelog entry concerning filepath buffer | 24 October 2016, 23:36:22 UTC |
321843d | Constantine Khroulev | 24 October 2016, 23:05:18 UTC | datetime.__sub__(): check calendars. | 24 October 2016, 23:05:41 UTC |
3694c5b | Constantine Khroulev | 24 October 2016, 21:25:26 UTC | Update Cython-generated C code. | 24 October 2016, 21:25:26 UTC |
4e4326f | Constantine Khroulev | 24 October 2016, 21:21:31 UTC | Merge branch 'master' into datetime-arithmetic | 24 October 2016, 21:21:31 UTC |
26645dd | Constantine Khroulev | 24 October 2016, 21:03:55 UTC | The 360_day calendar *has* year zero. | 24 October 2016, 21:03:55 UTC |
ac0b530 | Constantine Khroulev | 24 October 2016, 21:03:03 UTC | Restore Python 2.6 compatibility. | 24 October 2016, 21:03:03 UTC |
c59bd40 | Constantine Khroulev | 24 October 2016, 20:31:21 UTC | Choose one of the allowed calendar names. Also: use static typing in a few more places. | 24 October 2016, 20:31:21 UTC |
10009c6 | Constantine Khroulev | 24 October 2016, 20:30:23 UTC | Remove total_seconds() (we don't need it any more). | 24 October 2016, 20:30:23 UTC |
832cfa1 | Constantine Khroulev | 24 October 2016, 20:28:33 UTC | Fix a Python-2.6-compatibility workaround. The old code checked until the first exception, i.e. only the line d1.year = 1999 was processed. | 24 October 2016, 20:28:33 UTC |
2e3d1c3 | Constantine Khroulev | 24 October 2016, 20:26:53 UTC | Improve regression test coverage. The new netcdftime.datetime code has 100% line coverage (checked using coverage.py). | 24 October 2016, 20:26:53 UTC |
8773adb | Jeff Whitaker | 24 October 2016, 19:00:17 UTC | Merge pull request #591 from Unidata/issue590 docstring updates to clarify resolution of time calculations | 24 October 2016, 19:00:17 UTC |
48db4e9 | Constantine Khroulev | 24 October 2016, 16:21:04 UTC | A bit of cleanup. | 24 October 2016, 16:47:58 UTC |
ca602c7 | Constantine Khroulev | 24 October 2016, 16:12:44 UTC | Remove datetime.add_timedelta(). __add__() and __sub__() should work now. | 24 October 2016, 16:47:58 UTC |
2608e5d | Constantine Khroulev | 21 October 2016, 22:20:42 UTC | Implemented add_timedelta for all supported calendars. Needs testing. | 24 October 2016, 16:47:58 UTC |
bdfaffb | Constantine Khroulev | 20 October 2016, 19:11:18 UTC | Use static typing in some places. | 24 October 2016, 16:47:58 UTC |
dec854f | Constantine Khroulev | 19 October 2016, 18:53:56 UTC | Update regression tests. | 24 October 2016, 16:47:58 UTC |
f58d5e0 | Constantine Khroulev | 19 October 2016, 18:53:32 UTC | Fix __repr__(), implement __str__(). | 24 October 2016, 16:47:58 UTC |
d4bb303 | Constantine Khroulev | 18 October 2016, 05:55:23 UTC | Implement proper datetime comparison. Add a regression test. | 24 October 2016, 16:47:58 UTC |
35a3afa | Constantine Khroulev | 18 October 2016, 05:29:56 UTC | Whitespace changes. | 24 October 2016, 16:47:57 UTC |
9718392 | Constantine Khroulev | 18 October 2016, 04:08:10 UTC | Rename a variable to avoid a name clash. | 24 October 2016, 16:47:57 UTC |
16c2441 | Constantine Khroulev | 18 October 2016, 00:33:15 UTC | Remove superfluous "print" calls. | 24 October 2016, 16:47:57 UTC |
9fc4403 | Constantine Khroulev | 18 October 2016, 00:31:44 UTC | Add a pickling test. | 24 October 2016, 16:47:57 UTC |
e17fe53 | Jonathan Helmus | 21 October 2016, 16:17:01 UTC | BUG: Increase buffer used to store filepath Increase the buffer used to store the filepath by one character to accommodate the NULL character copied by the nc_inq_path function. closes #585 | 21 October 2016, 16:17:01 UTC |
de1db72 | Constantine Khroulev | 18 October 2016, 00:28:54 UTC | Rename the test case. All tests in this class test datetime methods, so this name seems appropriate to me. | 18 October 2016, 06:31:37 UTC |
1e527ed | Constantine Khroulev | 18 October 2016, 00:14:39 UTC | Update datetime.__reduce__(). | 18 October 2016, 06:31:36 UTC |
56d0216 | Constantine Khroulev | 17 October 2016, 23:50:06 UTC | Use "cdef readonly" and remove the "calendar" property. | 18 October 2016, 06:31:36 UTC |
1aef278 | Constantine Khroulev | 17 October 2016, 23:45:34 UTC | Return NotImplemented instead of raising TypeError. | 17 October 2016, 23:45:34 UTC |
1c4d742 | Constantine Khroulev | 17 October 2016, 23:35:53 UTC | Implement total_seconds. datetime.timedelta.total_seconds() is new in Python 2.7. | 17 October 2016, 23:42:14 UTC |
7aeda7c | Constantine Khroulev | 17 October 2016, 23:05:25 UTC | Add _netcdftime.c (?) I'm doing this to silence the following error from Travis CI: gcc: error: netcdftime/_netcdftime.c: No such file or directory | 17 October 2016, 23:05:25 UTC |
e306c8a | Constantine Khroulev | 17 October 2016, 23:02:22 UTC | Specify calendar as a keyword argument. | 17 October 2016, 23:02:22 UTC |
5309a96 | Constantine Khroulev | 17 October 2016, 21:25:05 UTC | Implement netcdftime.datetime.replace(**kwargs). | 17 October 2016, 21:39:22 UTC |
97c4323 | Constantine Khroulev | 17 October 2016, 19:43:23 UTC | Add regression tests. | 17 October 2016, 21:39:22 UTC |
a434c13 | Constantine Khroulev | 17 October 2016, 19:23:08 UTC | Use cdef and cpdef to speed things up a bit. | 17 October 2016, 21:39:22 UTC |
38036ce | Constantine Khroulev | 17 October 2016, 18:45:20 UTC | Implement datetime and timedelta arithmetic. | 17 October 2016, 21:39:22 UTC |
a0a0266 | Constantine Khroulev | 17 October 2016, 18:12:33 UTC | netcdftime/{netcdftime.py => _netcdftime.pyx} | 17 October 2016, 18:12:33 UTC |
d6bdb96 | Jeffrey Whitaker | 05 October 2016, 13:01:54 UTC | update | 05 October 2016, 13:01:54 UTC |
c02014d | Jeffrey Whitaker | 05 October 2016, 12:57:41 UTC | regenerate docs | 05 October 2016, 12:57:41 UTC |
9d13f3d | Jeffrey Whitaker | 05 October 2016, 12:54:24 UTC | regenerate C | 05 October 2016, 12:54:24 UTC |
6bcc4a8 | Jeffrey Whitaker | 05 October 2016, 12:53:14 UTC | docstring updates to clarify accuracy of time calculations. | 05 October 2016, 12:53:14 UTC |
97f1515 | Jeff Whitaker | 22 September 2016, 02:17:48 UTC | Merge pull request #586 from ckhroulev/issue-584 360_day and 365_day calendars: support negative Julian days. See #584. | 22 September 2016, 02:17:48 UTC |
3f56a2d | Constantine Khroulev | 21 September 2016, 22:49:50 UTC | Restore Python 3.x compatibility. | 21 September 2016, 22:49:50 UTC |
261b054 | Constantine Khroulev | 21 September 2016, 22:29:42 UTC | Add a Changelog entry. | 21 September 2016, 22:29:42 UTC |
4c31823 | Constantine Khroulev | 21 September 2016, 22:18:13 UTC | Add regression tests. | 21 September 2016, 22:22:06 UTC |
29af677 | Constantine Khroulev | 16 September 2016, 19:35:19 UTC | 360_day and 365_day calendars: support negative Julian days. In 360-day and 365-day calendars all years have the same length, so it is easy to shift the input by an integer number of years without affecting the date within the year. We can then use an algorithm that requires non-negative Julian days and later shift the year back to its original value. The day of the week requires a correction: in the 365-day calendar every year of the "shift" changes the day of the week number by 1 (365 mod 7 = 1). The old code for the 360-day calendar does not compute the day of the week, so no correction is applied. | 16 September 2016, 19:35:19 UTC |
c0810ba | Jeff Whitaker | 09 August 2016, 17:48:06 UTC | Merge pull request #577 from Unidata/valid_minmax use valid_min/valid_max/valid_range when creating mask (issue #576) | 09 August 2016, 17:48:06 UTC |
0d88852 | Jeffrey Whitaker | 04 August 2016, 19:46:52 UTC | treat case when fill_value = 0 | 04 August 2016, 19:46:52 UTC |
1535f54 | Jeffrey Whitaker | 03 August 2016, 17:24:20 UTC | empty commit to trigger Appveyor build | 03 August 2016, 17:24:20 UTC |
f206d85 | Jeffrey Whitaker | 03 August 2016, 14:03:32 UTC | update | 03 August 2016, 14:03:32 UTC |
b985ef5 | Jeffrey Whitaker | 03 August 2016, 12:35:23 UTC | make sure _FillValue has correct endian-ness | 03 August 2016, 12:35:23 UTC |