Closed ldecicco-USGS closed 2 years ago
This seems reasonable -- Thinking about how it would have been done if this "no errors ever on web service calls" had been the initial design, I feel like it would have ultimately done what you are doing if in a little less "bolt on" way. I'd say :shipit:
Daw...I haven't seen the ship-it squirrel in forever!
We've been pinged for not failing gracefully on our examples in CRAN (this is the 2nd time). CRAN doesn't want the functions to
stop
orerror
if the server returns an error message. So...I switched thegetWebService
function to usehttr::message_for_status
(instead ofhttr::stop_for_status
), and return aNULL
. This is consistent with what @dblodgett-usgs is doing withsbtools
.@wdwatkins or @dblodgett-usgs ...if one of you could glance at the code to see if it looks good to you, that would be great. I actually got to test it quite extensively yesterday because WQP was failing all over the place (someone was overloading them).
We need to resubmit to CRAN in the next couple of days to avoid having it taken down.