STINNER Victor <victor.stinner@haypocalc.com> added the comment: I think that threading._get_ident() is more reliable than threading.current_thread().ident because Thread.ident can be None in some cases. ---------- _______________________________________ Python tracker <report@bugs.python.org> <http://bugs.python.org/issue12028> _______________________________________

STINNER Victor <victor.stinner@haypocalc.com> added the comment: The threading API changed in Python 3 with the PEP 3108, especially and issue #2875. Extract of the PEP: "thread [done] * People should use 'threading' instead. * Rename 'thread' to _thread. * Deprecate dummy_thread and rename _dummy_thread. * Move thread.get_ident over to threading. * Guido has previously supported the deprecation [13]." The issue was closed by the commit [7fe3a8648ce2], "Issue #2875: Rename thread to _thread and dummy_thread to _dummy_thread." ---------- nosy: +brett.cannon, georg.brandl _______________________________________ Python tracker <report@bugs.python.org> <http://bugs.python.org/issue12028> _______________________________________

STINNER Victor <victor.stinner@haypocalc.com> added the comment:

In which case can it be None?

Oh, I misunderstood threading.py. current_thread().ident cannot be None. During the bootstrap of a thread, Thread._ident is None, but current_thread().ident is not None because current_thread() creates a dummy thread object having the right identifer. This dummy object is destroyed at soon as the Thread object is started (see Thread._bootstrap_inner). current_thread().ident is a little bit "suboptimal" because it gets the identifier of the thread, it reads the corresponding thread object, and then it gets the identifier of the thread object. def current_thread(): ... return _active[_get_ident()] ... ---------- _______________________________________ Python tracker <report@bugs.python.org> <http://bugs.python.org/issue12028> _______________________________________

Éric Araujo <merwok@netwok.org> added the comment: I find it strange that you propose to make a function public and remove its doc :) Is a negative missing in the bug title? ---------- nosy: +eric.araujo _______________________________________ Python tracker <report@bugs.python.org> <http://bugs.python.org/issue12028> _______________________________________

STINNER Victor <victor.stinner@haypocalc.com> added the comment: threading_get_ident.patch: make get_ident() public, replace threading._get_ident() by threading.get_ident(). According to this patch, get_ident() function *is* used: it is used by the logging and reprlib modules (and many tests). My patch avoids the usage of the low-level module, _thread_, in logging and reprlib. I adapted signal.pthread_kill() documentation to replace threading.current_thread().ident by threading.get_ident(). ---------- keywords: +patch Added file: http://bugs.python.org/file22089/threading_get_ident.patch _______________________________________ Python tracker <report@bugs.python.org> <http://bugs.python.org/issue12028> _______________________________________

Roundup Robot <devnull@devnull> added the comment: New changeset 6387aa6b496d by Victor Stinner in branch 'default': Issue #12028: fix NEWS entry http://hg.python.org/cpython/rev/6387aa6b496d ---------- _______________________________________ Python tracker <report@bugs.python.org> <http://bugs.python.org/issue12028> _______________________________________