Version ID bug #22
Description
I just noticed that I overlooked an difference introduced by Preston while
calculating the version.
We decided the version of a Dataset would be the sum of all involved node
versions for a dataset. So for example, a dataset would have the version id
calculated of the sum of the dataset "vid", plus the related people "vid", plus
the related variable "vid"s, and so on. Preston just changed this to get all
collapsed in one case, the packageID, with the "vid" of the dataset. This is
not correct, as a dataset changes when any of its components changes, including
all related nodes.
The temporary fix would involve just comment the way package_id is calculated
at the end of the "eml_variables"
Original issue reported on code.google.com by inigo.sa...@gmail.com on 9 May 2012 at 10:15