[Python-Dev] Remove own implementation for thread-local storage

[Antoine Pitrou]

Also note that C11, MSVC and some Unix C compilers have built-in support
for thread-local variables.  Example:
https://github.com/numba/numba/blob/master/numba/_random.c#L114-L119

Regards

Antoine.


On Mon, 3 Jul 2017 10:07:06 +0200
Victor Stinner <victor.stinner at gmail.com> wrote:

> I'm in favor of removing it. I know that it confused people many
> times, they look at this fallback and found an issue, whereas I'm not
> aware of any platform using this fallback anymore.
> 
> Can you please write a PR just to remove this fallback? We can merge
> it and then check buildbots :-) So in the worst case, we can revert
> it.
> 
> Victor
> 
> 2017-07-03 9:32 GMT+02:00 Masayuki YAMAMOTO <ma3yuki.8mamo10 at gmail.com>:
> > Hi, python-dev.
> >
> > I'd propose removing code which I think out-of-date.
> > CPython has provided the own implementation for thread-local storage (TLS)
> > on Python/thread.c, it's used in the case which a platform has not supplied
> > native TLS.  However, currently all supported platforms (NT and pthreads)
> > have provided native TLS and defined the Py_HAVE_NATIVE_TLS macro with
> > unconditional in any case.
> > If the code is removed, the new TLS API for PEP 539 won't have to care the
> > reinitialization of the thread keys managed by the interpreter (i.e.
> > PyThread_ReInitTLS function has been working for own implementation and will
> > be no longer necessary for new API).  Does anyone have a reason we should
> > keep it?
> >
> > Regards,
> > Masayuki
> >
> > _______________________________________________
> > Python-Dev mailing list
> > Python-Dev at python.org
> > https://mail.python.org/mailman/listinfo/python-dev
> > Unsubscribe:
> > https://mail.python.org/mailman/options/python-dev/victor.stinner%40gmail.com
> >