[Top][All Lists]

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

Re: [MIT-Scheme-devel] Symmetric MultiProcessing

From: Alex Shinn
Subject: Re: [MIT-Scheme-devel] Symmetric MultiProcessing
Date: Mon, 20 Jan 2014 10:04:31 +0900

On Mon, Jan 20, 2014 at 4:16 AM, Taylor R Campbell <address@hidden> wrote:
   Date: Sun, 19 Jan 2014 10:58:15 -0700
   From: Matt Birkholz <address@hidden>

   > From: Chris Hanson <address@hidden>
   > Date: Fri, 17 Jan 2014 22:40:01 -0800
   > R7RS defines parameters, which we'll eventually need to implement anyway
   > for compliance.  Converting the system's bindings over to parameters is
   > probably the right thing for a variety of reasons.

   Then I will stick it to master, one binding per commit.

Let's keep it in a branch until we get 9.2 out, at least.

   Comply with R7RS?  Does it not require multiple-value returns?

Probably, but it's a bug that we don't implement them right.

Lying around somewhere I might have a local branch with a stab at
implementing them with zero overhead for the single-return,
single-receive case by adding a return code for the multiple-receive
case.  In principle this should work, but integrating it with compiled
code made it tricky.

It's pretty trivial to support MVs entirely in user-space
by wrapping them in a special object, at zero overhead
for the single value case.

You do need to add an extra check for the wrapped type
in call/cc, but there's no compiler support required.


reply via email to

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