Re: [Chicken-hackers] [PATCH(5)] Remove srfi-18 and srfi-69

[Felix Winkelmann]

From: Christian Kellermann <address@hidden>
Subject: Re: [Chicken-hackers] [PATCH(5)] Remove srfi-18 and srfi-69
Date: Fri, 12 Sep 2014 12:15:10 +0200

> * Felix Winkelmann <address@hidden> [140912 12:11]:
>> From: Peter Bex <address@hidden>
>> Subject: Re: [Chicken-hackers] [PATCH(5)] Remove srfi-18 and srfi-69
>> Date: Fri, 12 Sep 2014 11:40:17 +0200
>> 
>> > On Fri, Sep 12, 2014 at 10:39:59AM +0200, Felix Winkelmann wrote:
>> >> Hello!
>> >> 
>> >> 
>> >> This patch removes support for srfi-18 and srfi-69. I had to remove
>> >> some tests as well, specifically those that use threads.
>> > 
>> > This seems like a bad idea.  These tests are there because we've had
>> > very hairy problems with the scheduler in the past.  Is it possible to
>> > keep the tests, using lower-level primitives?
>> 
>> I have moved the last remaining tests (signal-tests.scm) into the
>> srfi-18 test suite, so all tests remain.
> 
> I guess the underlying issue is the question: "How do we test the
> scheduler now?" And I think the answer to that is simply new
> tests. I am in the process of preparing a scheduler module, so maybe
> that also will make things a bit easier.

What is the intention behind a scheduler module? There is only one
procedure ("##sys#schedule"), which should definitely not be exported.


felix