[P4-design] undefined reads

Mihai Budiu mbudiu at vmware.com
Mon Jun 19 14:18:05 EDT 2017

The C guys have a lot of constraints for backwards compatibility.
We can do better.


From: P4-design [mailto:p4-design-bounces at lists.p4.org] On Behalf Of Vladimir Gurevich
Sent: Saturday, June 17, 2017 1:10 PM
To: Nate Foster <jnfoster at cs.cornell.edu>
Cc: p4-design <p4-design at lists.p4.org>
Subject: Re: [P4-design] undefined reads

Hi Nate,

This is a truly fascinating read! Thanks for the pointer.

For me, the main takeaway, though, is that perhaps we should take it easy :)  If C guys couldn't figure it out in 28 years (counting from C89 to be more than fair) or more, we can probably wait a year or two so that we gain some experience :) In fact, I was just about to make a comment on another thread that it is not clear to me how well C spec defines these undeterminate (I like this term from the article) behaviors.

I think the general statements Andy is proposing a probably already better than what you can find in many other places.


On Sat, Jun 17, 2017 at 7:06 AM, Nate Foster <jnfoster at cs.cornell.edu<mailto:jnfoster at cs.cornell.edu>> wrote:
Hi folks,

We've had a lot of past discussion (and some ongoing discussion) about undefined values / behavior. Here is a nice ACM Queue that outlines some of these issues in the context of C. I think it's worth a read.



P4-design mailing list
P4-design at lists.p4.org<mailto:P4-design at lists.p4.org>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.p4.org/pipermail/p4-design_lists.p4.org/attachments/20170619/2253392f/attachment-0002.html>

More information about the P4-design mailing list