Alex Russell

Hey Ben, I understand your concerns, both substantive and implied ("as a Googler..."). There are a few things I ask you to consider and which swayed me: - multiple platforms...

I don't like `.lock()`/`.unlock()` either. The current impl in the polyfill _might_ do the right thing in this case on Chrome/IE as I'm reusing transactions where possible...although it looks like...

@sicking : There's nothing wrong with calling `oncomplete` later than end-of-txn; the issue here is that FF calls back into user code that can still access the txn before that...

I like this a lot better than the add/remove churn. We should do this.

I'm absolutely for Apache or BSD. We'll need someone to own the rights. Lets move this to the overconstrained mailing list. On Apr 3, 2012, at 5:37 PM, Thomas Aylott...

https://groups.google.com/forum/?fromgroups#!forum/overconstrained On Apr 6, 2012, at 3:15 PM, Thomas Aylott wrote: > Where is that? > > On Apr 3, 2012, at 9:50 AM, Alex [email protected] wrote: > > >...

Hrm...the test you're adding doesn't work for me after I manually merge. What'd I miss?

Thanks.

The disjunctive constraints paper is here: http://www.cs.washington.edu/research/constraints/solvers/disjuncts.pdf

Note that the API and notification format should be broadly compatible with the Object.observe() straw-man: http://wiki.ecmascript.org/doku.php?id=strawman:observe