A reader has reported an error (and I'm pretty sure he's correct - see below) in the article at:

http://www.python.org/doc/essays/threads.html

This article isn't in the website svn and so I can't just correct it. I assume it is an old-and-no-longer-maintained article. The article itself starts with "This is an unfinished draft!".

I think the best course of action is just to delete it from the website. Anyone disagree?

All the best,

Michael

-------- Original Message --------
Subject:        A little mistake, Agustin
Date:   Fri, 17 Dec 2010 19:00:59 -0500
From:   [email protected]
To:     [email protected]




Hello guys,

I am computer engineer with big experience developing in C/C++.
First of all thanks for your work, I think it is a great contribution.
I was reading this page http://www.python.org/doc/essays/threads.html
And in my opinion the following paragraph contain a tiny mistake.
The words underlain are in reverse order, so will be [.... from locked to unlocked ..... ]

...
A lock is an object with two states: locked and unlocked. Initially, it is in the unlocked state. It has two methods: acquire() and release(). The acquire() method changes the lock from the unlocked into the locked state. This is like setting a semaphore to unsafe and entering the stretch of railroad tracks guarded by the semaphore. The release() method changes the lock from _unlocked to locked_; it is like leaving the stretch and resetting the semaphore to safe. When acquire() is called on a lock that is already locked, the operation *blocks* until another thread invokes the release() method. This is like a train waiting to enter the stretch until the semaphore signals safe. It is illegal to call release() when the lock is unlocked. This would be like a train leaving the guarded stretch without having set the semaphore to unsafe on entering -- a crash may occur!
...

Thanks a lot.
I have some time, so I would like to collaborate with the project in any way.
Agustin Fuentes.
_______________________________________________
pydotorg-www mailing list
[email protected]
http://mail.python.org/mailman/listinfo/pydotorg-www

Reply via email to