Like the old joke goes, I was wrong to say I was wrong. ... Yesterday's problem was not browser caching. I really was seeing different pull results from different instances of my server code. Turns out that the real issue is that I'm using the Vase 0.9.1-SNAPSHOT, and the behavior changed two weeks ago, after I had last refreshed it on one of my machines. The change occurred in one of the three commits on April 6-7, and results in a pull query returning one less level of wrapping. The new behavior seems more correct to me, but is an incompatible change. I'll file an issue with the details in a few minutes. @mtnygard

https://github.com/cognitect-labs/vase/issues/71