[libstdc++ PATCH/RFC] problem with pthread4 testcase

[Jason R Thorpe]

On Mon, Oct 28, 2002 at 10:43:24PM -0600, Loren James Rittle wrote:

 > NetBSD has a *brand new* thread implementation designed for deployment
 > in 2002 that doesn't even support concurrency on a single processor
 > machine with a special setup call?  Are you pulling my leg?

I'll go ahead and withdraw this patch ... I'm discussing with Nathan
what should be done about it:

	1. In response to your message, I implemented Solaris-compatible
	   pthread_{get,set}concurrency() in our pthread library (since
	   it is not merged into the mainline yet, I have the luxury of
	   making changes to it :-).

	   This change makes the test pass, as expected.

	2. Do we just bite the bullet and let bad pthreads programming
	   win, and just default to timeslicing by default.

	   Yes, I believe that any application that depends on timeslicing
	   being enabled by default is fundamentally broken.  But that's not
	   a debate that we should really have here :-)

Once we figure out what we're going to do about it, I'll either submit
another patch (case 1) or let it quietly drop on the floor (case 2).

-- 
        -- Jason R. Thorpe <thorpej@wasabisystems.com>