On Thu, Sep 17, 2009 at 2:10 PM, Ian Dees <[email protected]> wrote: > On Thu, Sep 17, 2009 at 5:16 AM, Jonathan Bennett > <[email protected]> wrote: >> Ian Dees wrote: >> >>> Perhaps the API >>> should ignore (i.e. not increase rev #'s) changes that don't actually >>> change anything? >> >> The idea isn't a bad one, but the implementation would need to be slick >> as to avoid bogging down the server -- it would double the number of DB >> queries per upload. > > It shouldn't have to perform more DB queries. The changeset already > has to be in memory (or at least available) at the time of changeset > close. Before the changeset is actually applied to the database, some > sort of collation could occur which would knock out all node changes > if they don't result in any change.
Changesets have nothing to do with it, especially changeset closing (when the close happens all the data is already committed). It's at the node/way/relation update level that that would be needed. Dave _______________________________________________ dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/dev
