In the current DO dump, geometries have IDs like us.census.tiger.puma, while in the new dump they look like us.census.tiger.puma_2015. Deploying the new dump will break all usages which have saved IDs, e.g: all Builder analysis using datasets from US, AU or CA.
This was introduced when adding the new 2016 data (conversation) because there was a check added during the timespan project that broke. We should have fixed the check (and maybe some code) instead and not change the IDs. In fact, allowing this was the main purpose of the timespans project.
This prevents us from deploying any changes to DO until its fixed.
In the current DO dump, geometries have IDs like
us.census.tiger.puma, while in the new dump they look likeus.census.tiger.puma_2015. Deploying the new dump will break all usages which have saved IDs, e.g: all Builder analysis using datasets from US, AU or CA. This was introduced when adding the new 2016 data (conversation) because there was a check added during the timespan project that broke. We should have fixed the check (and maybe some code) instead and not change the IDs. In fact, allowing this was the main purpose of the timespans project. This prevents us from deploying any changes to DO until its fixed.