This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author vstinner
Recipients gvanrossum, karamanolev, michaelg, r-englund, vstinner, yselivanov
Date 2015-03-14.08:29:57
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <CAMpsgwaayjRPhkYDp6aB3XieF32CiuZKLo1rQ+QzHL62PBC-Wg@mail.gmail.com>
In-reply-to <CAP7+vJKods0vYghY9Xb2Ay7-eVijjnBnLyzM3svd92--OXRHWw@mail.gmail.com>
Content
It looks like you are running your app on Windows. Are you using the
proactor event loop? I guess yes since you have more than 1000 clients and
select() is limited to 500 sockets.

In Python 3.4.3, I fixed a *lot* of crashes and race conditions in the
proactor event loop. There are maybe more race conditions. Are you running
subprocesses with asyncio?

Le 14 mars 2015 05:42, "Guido van Rossum" <report@bugs.python.org> a écrit :
> PyObject_GC_Malloc(unsigned __int64 basicsize=4046951880)
>
> That's nearly 4 GB. I somehow doubt your app is actually trying to
allocate
> that much memory

The caller is PyList_New(0).

When using a debugger on an optimized build, you should not trust the
debugger. For example, variables can be stored in registers. Displaying
values in the old frames can just display junk. In gdb, you see
<optimized>. The fast call convention uses registers to pass first
parameters to functions. I'm not sure that debuggers handle this case
correctly. Etc.
History
Date User Action Args
2015-03-14 08:29:58vstinnersetrecipients: + vstinner, gvanrossum, yselivanov, karamanolev, r-englund, michaelg
2015-03-14 08:29:58vstinnerlinkissue23187 messages
2015-03-14 08:29:57vstinnercreate