<div dir="ltr"><div class="gmail_extra">On Thu, Apr 4, 2013 at 1:22 PM, Richard Newman <span dir="ltr"><<a href="mailto:rnewman@mozilla.com" target="_blank">rnewman@mozilla.com</a>></span> wrote:<br><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
If promises resolve on the same tick *now*, then people will make their code work in that case. If some other part of the tree changes to resolve on a subsequent tick -- and this encapsulation of asynchrony is part of the value of promises -- then a distant caller could be subtly broken because of a timing inversion. In the worst case that assumption isn't even tested, and we don't know it breaks.<br>
</blockquote></div><br></div><div class="gmail_extra">Yeah, this is the essence of my concern: my concern isn't that we're asking coders to deal with nondeterministic behavior. We deal with that all the time. It's that we're asking them to deal with "deterministic, for now." behavior.<br>
<br>-dave<br></div></div>