Commit Graph

100 Commits

Author SHA1 Message Date
Victor Stinner ae6cd7cfda
bpo-37205: time.time() cannot fail with fatal error (GH-23314)
time.time(), time.perf_counter() and time.monotonic() functions can
no longer fail with a Python fatal error, instead raise a regular
Python exception on failure.

Remove _PyTime_Init(): don't check system, monotonic and perf counter
clocks at startup anymore.

On error, _PyTime_GetSystemClock(), _PyTime_GetMonotonicClock() and
_PyTime_GetPerfCounter() now silently ignore the error and return 0.
They cannot fail with a Python fatal error anymore.

Add py_mach_timebase_info() and win_perf_counter_frequency()
sub-functions.
2020-11-16 16:08:05 +01:00
Victor Stinner 3df5c68487
bpo-37205: time.perf_counter() and time.monotonic() are system-wide (GH-23284)
time.perf_counter() on Windows and time.monotonic() on macOS are now
system-wide. Previously, they used an offset computed at startup to
reduce the precision loss caused by the float type. Use
time.perf_counter_ns() and time.monotonic_ns() added in Python 3.7 to
avoid this precision loss.
2020-11-16 13:21:45 +01:00
Ronald Oussoren 41761933c1
bpo-41100: Support macOS 11 and Apple Silicon (GH-22855)
Co-authored-by:  Lawrence D’Anna <lawrence_danna@apple.com>

* Add support for macOS 11 and Apple Silicon (aka arm64)
   
  As a side effect of this work use the system copy of libffi on macOS, and remove the vendored copy

* Support building on recent versions of macOS while deploying to older versions

  This allows building installers on macOS 11 while still supporting macOS 10.9.
2020-11-08 10:05:27 +01:00
Minmin Gong f660567937
bpo-40650: Include winsock2.h in pytime.c, instead of a full windows.h (GH-20137) 2020-05-18 17:22:53 +01:00
Serhiy Storchaka eebaa9bfc5
bpo-38249: Expand Py_UNREACHABLE() to __builtin_unreachable() in the release mode. (GH-16329)
Co-authored-by: Victor Stinner <vstinner@python.org>
2020-03-09 20:49:52 +02:00
Michael Felt de6f38db48
bpo-39502: Fix 64-bit Python PyTime_localtime() on AIX (GH-18285)
Fix time.localtime() on 64-bit AIX to support years before 1902 and after 2038.
2020-02-07 18:56:16 +01:00
Benjamin Peterson f1c19031fd bpo-38068: Clean up gettimeofday configure logic. (GH-15775)
Assume gettimeofday exists and takes two arguments.
2019-09-10 03:37:59 -07:00
Victor Stinner 8709490f48
bpo-34373: Fix time.mktime() on AIX (GH-12726)
Fix time.mktime() error handling on AIX for year before 1970.

Other changes:

* mktime(): rename variable 'buf' to 'tm'.
* _PyTime_localtime():

  * Use "localtime" rather than "ctime" in the error message
    (specific to AIX).
  * Always initialize errno to 0 just in case if localtime_r()
    doesn't set errno on error.
  * On AIX, avoid abs() which is limited to int type.
  * EINVAL constant is now always available.
2019-04-09 19:12:26 +02:00
Michael Felt e2926b7248 bpo-34373: fix test_mktime and test_pthread_getcpuclickid tests on AIX (GH-8726)
* Fix test_mktime on AIX by adding code to get mktime to behave the
  same way as it does on other *nix systems
* Fix test_pthread_getcpuclickid in AIX by adjusting the test case
  expectations when running on AIX in 32-bit mode

Patch by Michael Felt.
2018-12-28 23:57:37 +10:00
luzpaz a5293b4ff2 Fix miscellaneous typos (#4275) 2017-11-05 15:37:50 +02:00
Victor Stinner c29b585fd4
bpo-31784: Implement PEP 564: add time.time_ns() (#3989)
Add new time functions:

* time.clock_gettime_ns()
* time.clock_settime_ns()
* time.monotonic_ns()
* time.perf_counter_ns()
* time.process_time_ns()
* time.time_ns()

Add new _PyTime functions:

* _PyTime_FromTimespec()
* _PyTime_FromNanosecondsObject()
* _PyTime_FromTimeval()

Other changes:

* Add also os.times() tests to test_os.
* pytime_fromtimeval() and pytime_fromtimeval() now return
  _PyTime_MAX or _PyTime_MIN on overflow, rather than undefined
  behaviour
* _PyTime_FromNanoseconds() parameter type changes from long long to
  _PyTime_t
2017-11-02 07:28:27 -07:00
Pablo Galindo 2c15b29aea bpo-31786: Make functions in the select module blocking when timeout is a small negative value. (#4003) 2017-10-17 17:14:41 +03:00
Victor Stinner bdaeb7d237 bpo-31773: _PyTime_GetPerfCounter() uses _PyTime_t (GH-3983)
* Rewrite win_perf_counter() to only use integers internally.
* Add _PyTime_MulDiv() which compute "ticks * mul / div"
  in two parts (int part and remaining) to prevent integer overflow.
* Clock frequency is checked at initialization for integer overflow.
* Enhance also pymonotonic() to reduce the precision loss on macOS
  (mach_absolute_time() clock).
2017-10-16 08:44:31 -07:00
Victor Stinner cba9a0c6de bpo-31773: time.perf_counter() uses again double (GH-3964)
time.clock() and time.perf_counter() now use again C double
internally.

Remove also _PyTime_GetWinPerfCounterWithInfo(): use
_PyTime_GetPerfCounterDoubleWithInfo() instead on Windows.
2017-10-12 08:51:56 -07:00
Victor Stinner 277c84067f Cleanup pytime.c (#3955)
* Move _PyTime_overflow() at the top
* Move assertion on numerator into _PyTime_ObjectToDenominator()
* PEP 7: add { ... } to if blocks
2017-10-11 08:11:38 -07:00
Victor Stinner a997c7b434 bpo-31415: Add _PyTime_GetPerfCounter() and use it for -X importtime (#3936)
* Add _PyTime_GetPerfCounter()
* Use _PyTime_GetPerfCounter() for -X importtime
2017-10-10 02:51:50 -07:00
Barry Warsaw b2e5794870 bpo-31338 (#3374)
* Add Py_UNREACHABLE() as an alias to abort().
* Use Py_UNREACHABLE() instead of assert(0)
* Convert more unreachable code to use Py_UNREACHABLE()
* Document Py_UNREACHABLE() and a few other macros.
2017-09-14 18:13:16 -07:00
Han Lee 829dacce4f bpo-26669: Fix nan arg value error in pytime.c (#3085)
* Fix #26669

* Modify NaN check function and error message

* Fix pytime.c when arg is nan

* fix whitespace
2017-09-08 16:05:05 -07:00
Benjamin Peterson a853a8ba78 bpo-31373: fix undefined floating-point demotions (#3396) 2017-09-07 11:13:59 -07:00
Antoine Pitrou bcaac8188b Revert "pytime: include winsock2, so we can have a complete timeval type (#3377)" (#3383)
This reverts commit 833860615b, as it broke Windows builds.
2017-09-06 13:31:09 +02:00
Benjamin Peterson 833860615b pytime: include winsock2, so we can have a complete timeval type (#3377) 2017-09-05 20:45:48 -07:00
haney c90e960150 bpo-30183: Fixes HP-UX cc compilation error in pytime.c (#1351)
* bpo-30183: Fixes HP-UX cc compilation error in pytime.c

HP-UX does not support the CLOCK_MONOTONIC identifier, and will fail to
compile:

    "Python/pytime.c", line 723: error #2020: identifier
    "CLOCK_MONOTONIC" is undefined
          const clockid_t clk_id = CLOCK_MONOTONIC;

Add a new section for __hpux that calls 'gethrtime()' instead of
'clock_gettime()'.

* bpo-30183: Removes unnecessary return
2017-06-21 20:18:21 +02:00
Alexander Belopolsky 3e7a3cb903 Issue #28148: Stop using localtime() and gmtime() in the time module.
Introduced platform independent _PyTime_localtime API that is similar
to POSIX localtime_r, but available on all platforms.  Patch by Ed
Schouten.
2016-09-28 17:31:35 -04:00
Benjamin Peterson af580dff4a replace PY_LONG_LONG with long long 2016-09-06 10:46:49 -07:00
Benjamin Peterson ed4aa83ff7 require a long long data type (closes #27961) 2016-09-05 17:44:18 -07:00
Victor Stinner e0a0afdf70 Merge 3.5 (pytime) 2015-11-10 12:11:52 +01:00
Victor Stinner c379ade1bb pytime.c: rename pygettimeofday_new() to pygettimeofday()
I forgot to rename it in my previous refactoring of pytime.c.
2015-11-10 12:11:39 +01:00
Serhiy Storchaka fad85aadb0 Issue #25558: Use compile-time asserts. 2015-11-07 15:42:38 +02:00
Victor Stinner b7a8af20ff Fix _PyTime_AsTimevalStruct_impl() on OpenBSD
On the x86 OpenBSD 5.8 buildbot, the integer overflow check is ignored. Copy
the tv_sec variable into a Py_time_t variable instead of "simply" casting it to
Py_time_t, to fix the integer overflow check.
2015-10-01 08:44:03 +02:00
Victor Stinner c29f399e7e Backout change 28d3bcb1bad6: "Try to fix _PyTime_AsTimevalStruct_impl() on
OpenBSD", I'm not sure that the change was really needed. I read the test
result of an old build because the OpenBSD was 100 builds late.
2015-09-30 22:50:12 +02:00
Victor Stinner 2bfed53b88 Try to fix _PyTime_AsTimevalStruct_impl() on OpenBSD
It looks like the check for integer overflow doesn't work on x86 OpenBSD 5.8.
2015-09-29 13:41:46 +02:00
Victor Stinner ec26f83f2e Issue #25155: Fix _PyTime_Divide() rounding
_PyTime_Divide() rounding was wrong: copy code from Python default which has
now much better unit tests.
2015-09-18 14:21:14 +02:00
Victor Stinner 9a8b177e60 Issue #25155: Add _PyTime_AsTimevalTime_t() function
On Windows, the tv_sec field of the timeval structure has the type C long,
whereas it has the type C time_t on all other platforms. A C long has a size of
32 bits (signed inter, 1 bit for the sign, 31 bits for the value) which is not
enough to store an Epoch timestamp after the year 2038.

Add the _PyTime_AsTimevalTime_t() function written for datetime.datetime.now():
convert a _PyTime_t timestamp to a (secs, us) tuple where secs type is time_t.
It allows to support dates after the year 2038 on Windows.

Enhance also _PyTime_AsTimeval_impl() to detect overflow on the number of
seconds when rounding the number of microseconds.
2015-09-18 13:36:17 +02:00
Victor Stinner 1e2b6882fc Issue #25155: Add _PyTime_AsTimevalTime_t() function
On Windows, the tv_sec field of the timeval structure has the type C long,
whereas it has the type C time_t on all other platforms. A C long has a size of
32 bits (signed inter, 1 bit for the sign, 31 bits for the value) which is not
enough to store an Epoch timestamp after the year 2038.

Add the _PyTime_AsTimevalTime_t() function written for datetime.datetime.now():
convert a _PyTime_t timestamp to a (secs, us) tuple where secs type is time_t.
It allows to support dates after the year 2038 on Windows.

Enhance also _PyTime_AsTimeval_impl() to detect overflow on the number of
seconds when rounding the number of microseconds.
2015-09-18 13:23:02 +02:00
Victor Stinner 51b9398444 pytime: oops, fix typos on Windows 2015-09-10 16:00:06 +02:00
Victor Stinner c60542b12b pytime: add _PyTime_check_mul_overflow() macro to avoid undefined behaviour
Overflow test in test_FromSecondsObject() fails on FreeBSD 10.0 buildbot which
uses clang. clang implements more aggressive optimization which gives
different result than GCC on undefined behaviours.

Check if a multiplication will overflow, instead of checking if a
multiplicatin had overflowed, to avoid undefined behaviour.

Add also debug information if the test on overflow fails.
2015-09-10 15:55:07 +02:00
Victor Stinner ff0ed3e71c New try to fix test_time.test_AsSecondsDouble() on x86 buildbots.
Use volatile keyword in _PyTime_AsSecondsDouble()
2015-09-10 13:25:17 +02:00
Victor Stinner 1efbebaac2 Try to fix test_time.test_AsSecondsDouble() on "x86 Gentoo Non-Debug with X 3.x" buildbot
Use volatile keyword in _PyTime_Round()
2015-09-10 11:48:00 +02:00
Victor Stinner 9c72f9b30a Fix test_time on Windows
* Filter values which would overflow on conversion to the C long type
  (for timeval.tv_sec).
* Adjust also the message of OverflowError on PyTime conversions
* test_time: add debug information if a timestamp conversion fails
2015-09-10 09:10:14 +02:00
Victor Stinner 3e2c8d84c6 test_time: rewrite PyTime API rounding tests
Drop all hardcoded tests. Instead, reimplement each function in Python, usually
using decimal.Decimal for the rounding mode.

Add much more values to the dataset. Test various timestamp units from
picroseconds to seconds, in integer and float.

Enhance also _PyTime_AsSecondsDouble().
2015-09-09 22:32:48 +02:00
Victor Stinner 9ae47dfbd9 pytime: add _PyTime_Round() helper to factorize code 2015-09-09 22:28:58 +02:00
Victor Stinner ce6aa749b4 Make _PyTime_RoundHalfEven() private again 2015-09-09 22:28:09 +02:00
Victor Stinner 7667f58151 Issue #23517: fromtimestamp() and utcfromtimestamp() methods of
datetime.datetime now round microseconds to nearest with ties going to nearest
even integer (ROUND_HALF_EVEN), as round(float), instead of rounding towards
-Infinity (ROUND_FLOOR).

pytime API: replace _PyTime_ROUND_HALF_UP with _PyTime_ROUND_HALF_EVEN. Fix
also _PyTime_Divide() for negative numbers.

_PyTime_AsTimeval_impl() now reuses _PyTime_Divide() instead of reimplementing
rounding modes.
2015-09-09 01:02:23 +02:00
Victor Stinner adfefa527a Issue #23517: Fix implementation of the ROUND_HALF_UP rounding mode in
datetime.datetime.fromtimestamp() and datetime.datetime.utcfromtimestamp().
microseconds sign should be kept before rounding.
2015-09-04 23:57:25 +02:00
Victor Stinner 5786aef382 Don't abuse volatile keyword in pytime.c
Only use it on the most important number. This change fixes also a compiler
warning on modf().
2015-09-03 16:33:16 +02:00
Victor Stinner 29ee6745af Enhance _PyTime_AsTimespec()
Ensure that the tv_nsec field is set, even if the function fails
with an overflow.
2015-09-03 16:25:45 +02:00
Victor Stinner 5a682c9ca5 Merge 3.5 (monotonic) 2015-09-03 00:15:23 +02:00
Victor Stinner 5ad5821d09 oops, rename pymonotonic_new() to pymonotonic()
I was not supposed to commit the function with the name pymonotonic_new(). I
forgot to rename it.
2015-09-03 00:14:58 +02:00
Victor Stinner c3c616c3d1 Issue #24707: Remove assertion in monotonic clock
Don't check anymore at runtime that the monotonic clock doesn't go backward.
Yes, it happens. It occurs sometimes each month on a Debian buildbot slave
running in a VM.

The problem is that Python cannot do anything useful if a monotonic clock goes
backward. It was decided in the PEP 418 to not fix the system, but only expose
the clock provided by the OS.
2015-09-03 00:13:46 +02:00
Victor Stinner 2ec558739e Issue #23517: datetime.timedelta constructor now rounds microseconds to nearest
with ties going away from zero (ROUND_HALF_UP), as Python 2 and Python older
than 3.3, instead of rounding to nearest with ties going to nearest even
integer (ROUND_HALF_EVEN).
2015-09-02 19:16:07 +02:00