2011/12/13 Richard Fairhurst <rich...@systemed.net>: > So there you go, full support for nested and ordered relations.
happy to see this (actually I discovered it myself just a few hours ago). > Martin, I've long since stopped expecting anything you say to be > founded in any form of reality. LOL, that's a good one. I was basing my comment on what you (or some other Richard Fairhurst?) were writing more then once to the lists, examples: 1. http://gis.638310.n2.nabble.com/Odd-Potlatch-2-rendering-td5892790.html RichardF: "Potlatch 2 will only draw multipolygons which consist of 1 outer way and 1+ inner ways, and where the tags are on the outer way. This multipolygon has several outer ways so will not be rendered properly. You could call this a bug but, to be honest, if people are going to invent insanely complicated relation structures (aka "advanced multipolygons") and completely go against existing OSM tagging practice (putting way tags on relations), I feel no obligation to waste a week of my life supporting it. It would be an exceptionally complex one to fix, as you'd potentially need to load extra data from the server if all the outer ways weren't in the current bbox. " 2. http://trac.openstreetmap.org/ticket/4048#comment:1 -> a few weeks ago you closed a trac ticket concerning this issue with the comment "won't fix". 3. http://gis.638310.n2.nabble.com/osm2pgsql-and-only-named-multipolygons-tt6858105.html Richard:"So for the tools I contribute to, principally P2 and an upcoming Ruby PDF renderer, I've taken a decision not to spend time on anything more than rudimentary multipolygon support (one outer, tags on outer way), rather than spending a month coding all-singing-all-dancing support and then have to rewrite it when the area tag comes along. YMalmost certainlyV. :) " and: "Frederik Ramm wrote: > While multipolygons with more than one outer *ring* are uncommon, > those with more than one outer *way* are heavily used where I live That's all right, everyone where you live uses JOSM anyway. :) The issue from my point of view is one of UI. I can't countenance a UI, or a render, that is any different for tagging shapes depending on whether they have holes in or not. So P2's tagging code would need to get some layer of indirection saying "if this is a multipolygon, the tag panel needs to deal with the relation tags, unless there are already tags on the way. Otherwise it needs to deal with the way tags". Meanwhile, the rendering code would need to render fills differently for items with more than one outer way (which is additionally complicated by the fact they might not all be loaded yet). That's far too much like brainache for me to want to code, especially when that code will need rewriting in a year or so. But if someone else who does come from a land where people use bonkers complicated multipolygons came up with a decent patch, I imagine we'd be very happy to accept it. " So in the end I am happy that someone recently coded the missing relation support for Potlatch2, but given your statements from previous discussions (as well as you closing relative tickets with "won't fix") wasn't really encouraging to think that this has been done in the past few weeks. Neither do I recall any announcement about this huge progress, nor would I have guessed this from the developer's comments on git: https://github.com/openstreetmap/potlatch2/commits/master?page=1 cheers, Martin _______________________________________________ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
