nothing fancy, but it works…

the output doesn't match the other quotes either .. maybe a to do .. or not?

prove to travis that my other commit actually works…

curl, curl_escape, and readJSON need to have their package prefixed to satisfy travis

travis didn't like `try( getQuote.IEX( 'vxx' ) )`. Can't test this locally so just trying to `source` the definition of `getQuote.IEX` in there…

Josh does not like `a->b`, so I switched them all to `b<-a`

upgrade travis build to Ubuntu 16, xenial xerus

was sourcing from *tests* folder. no wonder travis didn't like it

just figuring out travis…

delete & replace for travis…

copied dirk's `sample.travis.yaml` so I can edit other parts of this and get it to pass

travis is still on trusty, so…

looks like josh got an old version from dirk https://docs.travis-ci.com/user/languages/r/

it was in 
```structure(list(symbol = c("VXX", "SNAP"), price = c(12.785, 17.27),
               size = c(100L, 200L), time = c(1498075199.954, 1498075184.411),
               .Names = c("symbol", "price", "size", "time"),
               row.names = 1:2, class = "data.frame")```, not the very sensible calls to `try`

accidentally deleted some of  josh's tests

prefix with quantmod:: so travis finds these functions?

make that quantmod:::

save time on travis build (why is it taking so long??)

non-ascii characters messing with  R CMD check

travis doesn't like my `pandoc` instructions …

maybe travis doesn't like `expected.types`?

should speed up builds …  I used the wrong syntax before

want it to build on R 3.4 which is not supported in precise pangolin

found it…

there are hadleyverse dependencies in @jimhester and @craigcitro's `r` build @joshuaulrich doesn't want

switching back for the PR