[OSM-dev] Data corruption :) II

Stefan de Konink stefan at konink.de
Wed Nov 26 07:38:27 GMT 2008


On Wed, 26 Nov 2008, Frederik Ramm wrote:

> Stefan de Konink wrote:
> > - Dave: should I just automatically delete this mess with the constraint
> > that if the node would not have any members at all anymore. I will just
> > delete the entire structure?
>
> Maybe it would be worthwhile to retrieve the latest version from history
> instead; one should maybe pick and analyze some random cases and try to
> find out what the user wanted and what went wrong.

Seems for all random cases the `current' is correct.

> My suggestion for your current issue would be to leave it alone until we
> have API 0.6 in place, because with our current infrastructure you are
> fighting windmills - every day new inconsistencies are created. With any
> luck, that will be a thing of the past once we have full transactions
> with API 0.6; so AFTER that is active we can prepare one big "fix legacy
> inconsistencies" operation and remove them once and for all.

Not acceptable.


Stefan





More information about the dev mailing list