Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | oops thanks BlastRock NA for spotting this | Armin Rigo | 2019-01-31 | 1 | -1/+2 |
| | |||||
* | Issue #362 | Armin Rigo | 2019-01-28 | 1 | -67/+226 |
| | | | | | | | Add "thread canary" objects which are deallocated if the PyThreadState is explicitly deallocated by CPython. If the thread shuts down first, then instead the canary is inserted in a zombie list. In that case, we clear and delete properly the PyThreadState at the next occasion. | ||||
* | msvc support | Armin Rigo | 2018-08-08 | 1 | -1/+3 |
| | |||||
* | Issue #362 | Armin Rigo | 2018-03-15 | 1 | -1/+69 |
| | | | | | | Py_Finalize() will free any threadstate around, so in that case we must not call PyThreadState_Delete() any more on them from cffi_thread_shutdown(). | ||||
* | Push and pull at the mess | Armin Rigo | 2017-10-05 | 1 | -22/+24 |
| | |||||
* | More fun with _PyThreadState_Current becoming undefined in 3.7. | Armin Rigo | 2017-09-27 | 1 | -9/+17 |
| | |||||
* | Pull request #79 by xwang | Armin Rigo | 2017-03-28 | 1 | -6/+1 |
| | | | | | We can actually call PyThreadState_Delete(), which works without the GIL and seems to pass the same tests. | ||||
* | Change get_current_ts to always use _Py_atomic_load_relaxed when available. | Michael McGee | 2016-04-05 | 1 | -1/+1 |
| | |||||
* | Remove this field if it's not used | Armin Rigo | 2016-01-03 | 1 | -1/+3 |
| | |||||
* | Generalize the code for 'local_thread_state' and move it to | Armin Rigo | 2016-01-03 | 1 | -0/+134 |
misc_thread_common.h. |