alternate evaluation strategy with fewer glitches?

[philip k]

Hi,

I just read the thesis paper on Flapjax and was under the impression
that the breadth-first evaluation scheme presented there would
eliminate all glitches. However, the latest github commit message
suggests otherwise - can somebody explain the exact nature of the
glitches that the new evaluation strategy attempts to rule out?

Also a short explanation of how this is done would be very helpful.

Thanks a lot!

[Arjun Guha]

Haha. Yeah. The strategy presented in the Flapjax paper is sound, but it's tricky to implement correctly. I have a small variant of the Flapjax paper's strategy that's easier to implement and fixes some bugs. (It's still doing breath-first traversal, so nothing dramatically different.)

I'm traveling over the next two weeks. I'll do a writeup when I'm back.

Arjun

--
Flapjax home page: www.flapjax-lang.org
Flapjax list: groups.google.com/group/flapjax
Post: fla...@googlegroups.com
Unsubscribe: flapjax-u...@googlegroups.com