wasm-demo/demo/ermis-f/python_m/cur/0358

52 lines
1.9 KiB
Plaintext

From: eric at linux-hw.com (Eric Lee Green)
Date: Tue, 13 Apr 1999 21:32:23 GMT
Subject: threads
In-Reply-To: <1288080841-20523503@hypernet.com>
References: <1288080841-20523503@hypernet.com>
Message-ID: <Pine.LNX.4.04.9904131722450.14793-100000@england.local.net>
Content-Length: 1569
X-UID: 358
On Tue, 13 Apr 1999, Gordon McMillan wrote:
> Eric Lee Green wonders about Python threads:
>
> > My question is this: How "atomic" are basic Python variable
> > operations?
>
> This exact question was asked not long ago. My reply, and Guido's
> clarifications follow:
Ah. Okay. So Python is doing user-land threads and not Posix (OS-level)
threads?
The difference between the two doesn't make a difference for my
application, but I'd be curious to know in case I ever want to run this on
an SMP machine (where only OS-level threads are schedulable on multiple
processors).
In any event, I'm relieved to know that basic operations are safe. As long
as the semaphoring requirements are similar to "C", which appears to be
the case, I am just fine. I do see, from reading the documentation, that
file I/O operations are thread-safe. The only question then is whether
Python's socket operations are thread-safe if we're doing user-land
threads... I'd hate to bring all threads to a halt by reading from a
socket, because that was the whole point (to be able to do things
asynchronously).
Be a real bummer if I have to stop what I'm doing and start re-writing it
all into C++ :-(. (I am *NOT* fond of C++, it seems you have to spend eons
writing basic classes to do things that Python does out of the box before
you can even think about your application).
--
Eric Lee Green eric at linux-hw.com http://www.linux-hw.com/~eric
"People have grown used to thinking of computers as unreliable, and it
doesn't have to be that way." -- Linus