reimplement Matt's method from c83fe5e with the bug fixed
@mbjones I liked your method and wanted to see why it wasn't working - turns out the bind function you created made a list that looked like:rbind(source1), rbind(source2), rbind(source3), which was run sequentially and then just returned source3.
I corrected it so that it returns rbind(source1, source2, source3)
Your tests were passing since you didn't have any keys set (I'm guessing), so it was only running one source. GHA, which runs all of the sources, found the issue though.
@mbjones I liked your method and wanted to see why it wasn't working - turns out the bind function you created made a list that looked like:
rbind(source1), rbind(source2), rbind(source3)
, which was run sequentially and then just returnedsource3
.I corrected it so that it returns
rbind(source1, source2, source3)
Your tests were passing since you didn't have any keys set (I'm guessing), so it was only running one source. GHA, which runs all of the sources, found the issue though.