[OSM-talk] only partial loading of ways with JOSM?
Tom Hughes
tom at compton.nu
Thu May 10 14:49:46 BST 2007
In message <f1v70f$upo$1 at sea.gmane.org>
Gabriel Ebner <ge at gabrielebner.at> wrote:
>> I also have an other problem. Old (deleted or moved) nodes, are back
>> in some areas.
>> Unfortunately I don't have an example, because I removed them again.
>
> I saw them too. And I didn't like them. So I deleted them. And guess what
> I saw?
>
> upload to: http://www.openstreetmap.org/api/0.4/node/25267534
> got return: 410 with id 25267534
> upload to: http://www.openstreetmap.org/api/0.4/node/25267585
> got return: 410 with id 25267585
> upload to: http://www.openstreetmap.org/api/0.4/segment/21249501
> got return: 200 with id 21249501
> upload to: http://www.openstreetmap.org/api/0.4/segment/21249499
> got return: 200 with id 21249499
>
> Deleting the nodes apparently gives 410 (Gone), so they're probably really
> dead. But the segments, there I get 200 (Ok), so they've probably still
> been "visible" before.
Correct. Or at least that is what I'm seeing.
> Could it be that all these segments were still visible in the db, but we
> never saw them because the referenced deleted nodes? And that we're now
> seeing them because a bug sends nodes regardless of whether they're visible
> or not?
I don't think that such a bug exists with regard to nodes.
I think the change is specific to the case where a deleted node is
referenced by a non-deleted segment - obviously that shouldn't really
be possible but something has allowed it to happen it the past.
Before yesterday evening the API would not return a segment one or
both of whose ends was deleted. Now it does, and when it does so it
also returns both ends even if those nodes are actually deleted.
In some ways this is an improvement because it means people can see
the bogus segments in josm and delete them. Once that is done neither
they nor the nodes they reference should show up again.
Whether this change to the API was deliberate or not is not clear...
Tom
--
Tom Hughes (tom at compton.nu)
http://www.compton.nu/
More information about the talk
mailing list