[IPython-dev] iTorch - Weird issues with mkfifo on linux

[Soumith Chintala]

This is not a problem of blocking reads, or the heartbeat stopping, I have verified both those parts.
But I’ll try investigating more into if it’s a system-call/signaling weirdness.
Thanks,
Soumith

From: Thomas Kluyver <takowl at gmail.com<mailto:takowl at gmail.com>>
Reply-To: IPython developers list <ipython-dev at scipy.org<mailto:ipython-dev at scipy.org>>
Date: Friday, January 16, 2015 at 10:16 AM
To: IPython developers list <ipython-dev at scipy.org<mailto:ipython-dev at scipy.org>>
Subject: Re: [IPython-dev] iTorch - Weird issues with mkfifo on linux

On 15 January 2015 at 19:26, Soumith Chintala <soumith at fb.com<mailto:soumith at fb.com>> wrote:
One problem we oncovered is that on certain (or possibly all linux distros) ipython thinks our kernel died, even though it is alive and well.

FIFOs have some bizarre properties which might be related - for instance, opening a FIFO for reading will block until something else opens it for writing (the open() call itself blocks, not just attempts to read). Could it be that an unexpectedly blocking system call is stopping the heartbeat and causing IPython to think that your kernel is dead?

Thomas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/ipython-dev/attachments/20150116/37b89aa3/attachment.html>