-
-
Notifications
You must be signed in to change notification settings - Fork 474
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Remove support for old-style MPs #862
Conversation
c007c49
to
5ac4993
Compare
I had forgotten the part where tags are copied from ways to relations. That is gone now, too. I think that covers all old-style MP code. @pnorman, it would help if you could look over the tagtransform code, if the handling is still right |
In context of #865: does this mean that the concept of a "pending way" becomes obsolete, and could be removed as well?
|
There is no more pending step during import but it is still needed for updates, where changes on nodes/way need to be propagated upwards to ways/relations. |
bumping this for myself - haven't gotten to look at it with holidays, forest fires, job hunting, and other stuff. |
@pnorman Did you have the time to look at this? You seem really busy but it'd be awesome to have this I think (when I see how much ways need too be processed for France only... it's killing my server!). |
07c0a92
to
bad810a
Compare
Also removes support for superseding ways from polygons. The lua bindings still expect that the supersede table is present but will simply drop it.
Remove some of the tests that test old-style MP behaviour.
The tag type=multipolygon already indicates an area, no matter what the other tags say. Fixes osm2pgsql-dev#866.
bad810a
to
5768610
Compare
This has been lying around for a while now. I have some more refactoring changes pending and like to get this in to avoid merge conflicts. If there are no further objections by the end of the week, I'll just get it merged. |
Also removes support for superseding ways from polygons. The lua bindings still expect that the supersede table is present but will simply drop it.
Still todo:
I've done some performance tests as this PR moves all way processing to the mono-threaded first computation stage. Test were done on a Hetzner EX41-SSD (i7-7700, 32GB RAM, SSD in Raid0) with command:
The minimal cache is a bit of a worst case scenario as all node lookups go to the slightly slower flatnode file.
master:
This PR:
So this looks fairly promising and it would be ok to go ahead with dropping old-style MPs without improving the first-stage processing first.