Re: [Python-Dev] _PyThreadState_Current
Since it's a regression introduced in Python 3.5.1, I propose to introduce a new private function _PyThreadState_FastGet() to reintroduce the feature: https://bugs.python.org/issue26154 Using afunction instead of using directly the variable hides how atomic variables are implemented and so avoid compiler issues. Victor 2016-01-18 21:18 GMT+01:00 Maciej Fijalkowski: > Hi > > change in between 3.5.0 and 3.5.1 (hiding _PyThreadState_Current and > pyatomic.h) broke vmprof. The problem is that as a profile, vmprof can > really encounter _PyThreadState_Current being null, while crashing an > interpreter is a bit not ideal in this case. > > Any chance, a) _PyThreadState_Current can be restored in visibility? > b) can I get a better API to get it in case it can be NULL, but also > in 3.5 (since it works in 3.5.0 and breaks in 3.5.1) > > Cheers, > fijal > ___ > Python-Dev mailing list > Python-Dev@python.org > https://mail.python.org/mailman/listinfo/python-dev > Unsubscribe: > https://mail.python.org/mailman/options/python-dev/victor.stinner%40gmail.com ___ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com
Re: [Python-Dev] _PyThreadState_Current
seems to work thanks. That said, I would love to have PyThreadState_Get equivalent that would let me handle the NULL. On Mon, Jan 18, 2016 at 9:31 PM, Maciej Fijalkowskiwrote: > Good point > > On Mon, Jan 18, 2016 at 9:25 PM, Victor Stinner > wrote: >> Hum, you can try to lie and define Py_BUILD_CORE? >> >> Victor >> >> 2016-01-18 21:18 GMT+01:00 Maciej Fijalkowski : >>> Hi >>> >>> change in between 3.5.0 and 3.5.1 (hiding _PyThreadState_Current and >>> pyatomic.h) broke vmprof. The problem is that as a profile, vmprof can >>> really encounter _PyThreadState_Current being null, while crashing an >>> interpreter is a bit not ideal in this case. >>> >>> Any chance, a) _PyThreadState_Current can be restored in visibility? >>> b) can I get a better API to get it in case it can be NULL, but also >>> in 3.5 (since it works in 3.5.0 and breaks in 3.5.1) >>> >>> Cheers, >>> fijal >>> ___ >>> Python-Dev mailing list >>> Python-Dev@python.org >>> https://mail.python.org/mailman/listinfo/python-dev >>> Unsubscribe: >>> https://mail.python.org/mailman/options/python-dev/victor.stinner%40gmail.com ___ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com
Re: [Python-Dev] _PyThreadState_Current
Good point On Mon, Jan 18, 2016 at 9:25 PM, Victor Stinnerwrote: > Hum, you can try to lie and define Py_BUILD_CORE? > > Victor > > 2016-01-18 21:18 GMT+01:00 Maciej Fijalkowski : >> Hi >> >> change in between 3.5.0 and 3.5.1 (hiding _PyThreadState_Current and >> pyatomic.h) broke vmprof. The problem is that as a profile, vmprof can >> really encounter _PyThreadState_Current being null, while crashing an >> interpreter is a bit not ideal in this case. >> >> Any chance, a) _PyThreadState_Current can be restored in visibility? >> b) can I get a better API to get it in case it can be NULL, but also >> in 3.5 (since it works in 3.5.0 and breaks in 3.5.1) >> >> Cheers, >> fijal >> ___ >> Python-Dev mailing list >> Python-Dev@python.org >> https://mail.python.org/mailman/listinfo/python-dev >> Unsubscribe: >> https://mail.python.org/mailman/options/python-dev/victor.stinner%40gmail.com ___ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com