[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Chicken-users] Please do not drop 'thread-terminate!' from the SRFI

From: Vincent Manis
Subject: Re: [Chicken-users] Please do not drop 'thread-terminate!' from the SRFI 18 impl
Date: Fri, 8 Aug 2008 19:32:13 -0700

On 2008-Aug-8, at 16:30, Kon Lovett wrote:

My 2 cents.

The SRFI document is clear about the danger. The Chicken mail archive is clear about the danger. "Standards & Practices" is clear about the danger.

Who are we protecting?

Well, I spent several years teaching concurrency (along with many other topics) at a large multinational company whose software engineers overwhelmingly spoke C++, and the number of horrific travesties they committed would sicken anyone. (A standard response to race conditions was to insert a sleep call, for example. `Oh, you QA folks are seeing crashes on your test machines? We can't reproduce them on our development machines.')

We should keep in mind that the mailing list probably focuses on the high end of Chicken users, and that many programmers don't necessarily understand this stuff that well, even if they did take the obligatory 3rd year OS course where they used mutexes and condition variables.

My preference is the same as John Cowan's: kill it. If not, put warnings in the manual. I don''t find SRFI-18's warning (in the paragraph headed NOTE:) strong enough, as it uses the phrase `may be a problem', not `is virtually guaranteed to be a problem'. Further, I don't know what fraction of Chicken users actually go and read the SRFIs.

So, I'd say, `we're protecting that large group of programmers whom we would like to persuade that Chicken is a Good Thing'.

-- v

reply via email to

[Prev in Thread] Current Thread [Next in Thread]