[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: the state of the concurrency branch
From: |
Tom Tromey |
Subject: |
Re: the state of the concurrency branch |
Date: |
Tue, 27 Aug 2013 20:34:21 -0600 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/24.3 (gnu/linux) |
>> The basic issue is that only one thread can select on a given fd at a
>> time.
Stefan> Why?
To me it seemed like the simplest approach, not requiring changes
anywhere else in Emacs: if select says data is available, the thread can
proceed to read it and handle it. Maybe another choice is to make the
reading code robust to failures here; and also audit to make sure there
aren't any possible races (readers have to really read all the data when
given a chance or else you could have different threads reading
different bits from the fd, possibly processing them in the wrong
order).
This approach was also convenient since it enabled thread-locking of
processes, something I think is desirable anyway.
Tom
- Re: the state of the concurrency branch, (continued)
- Re: the state of the concurrency branch, Stefan Monnier, 2013/08/27
- Re: the state of the concurrency branch, Eli Zaretskii, 2013/08/28
- Re: the state of the concurrency branch, Tom Tromey, 2013/08/28
- Re: the state of the concurrency branch, Eli Zaretskii, 2013/08/28
- Re: the state of the concurrency branch, Stefan Monnier, 2013/08/28
- Re: the state of the concurrency branch, Tom Tromey, 2013/08/28
- Re: the state of the concurrency branch, Stefan Monnier, 2013/08/28
- Re: the state of the concurrency branch, Stefan Monnier, 2013/08/27
- Re: the state of the concurrency branch,
Tom Tromey <=
- Re: the state of the concurrency branch, Tom Tromey, 2013/08/27
- Re: the state of the concurrency branch, Eli Zaretskii, 2013/08/28