[OSM-dev] Recursive relations

2009-09-13 Thread Andrew M. Bishop
In OSM a relation can contain other relations but it seems that there is nothing to check that a relation doesn't contain itself. I can't think of any legitimate reason that it should be allowed though. For example relation 15852 has contained itself since version 108 which was created at the

Re: [OSM-dev] Recursive relations

2009-09-13 Thread Shaun McDonald
From a data and technical point of view this has always been possible in the API intentionally. It's just that no one has come up with a good use for it yet. Shaun On 13 Sep 2009, at 10:56, Andrew M. Bishop wrote: In OSM a relation can contain other relations but it seems that there is

Re: [OSM-dev] Recursive relations

2009-09-13 Thread Karl Guggisberg
This is probably something that the OSM editors should warn users about when they try to create such a relation. JOSM doesn't allow to create recursive relations, but it accepts them when they are present in the data. -- Karl -Ursprüngliche Nachricht- Von: dev-boun...@openstreetmap.org

Re: [OSM-dev] Recursive relations

2009-09-13 Thread Maarten Deen
Karl Guggisberg wrote: This is probably something that the OSM editors should warn users about when they try to create such a relation. JOSM doesn't allow to create recursive relations, but it accepts them when they are present in the data. JOSM has in the past accepted recursive relations

[OSM-dev] Wiki: Renaming the API pages

2009-09-13 Thread Lars Francke
Hi, I already asked this on the Talk page for the API (http://wiki.openstreetmap.org/wiki/Talk:API) but I don't know how many people will be reading that. * I'd like to rename (move) Protocol to API on the Wiki: OSM_Protocol_Version_0.6 - OSM_API_Version_0.6. API is the commonly used word and it

Re: [OSM-dev] Wiki: Renaming the API pages

2009-09-13 Thread Thomas Wood
+1 This has needed a tidyup for a long time. 2009/9/13 Lars Francke lars.fran...@gmail.com: Hi, I already asked this on the Talk page for the API (http://wiki.openstreetmap.org/wiki/Talk:API) but I don't know how many people will be reading that. * I'd like to rename (move) Protocol to API

[OSM-dev] This week's planet .osm

2009-09-13 Thread Richard Ive
Hey all, Has anyone had any issue with this weeks planet dump? 090909 I've tried to rum an osm2pgsql import twice now and it's falled over at Relation 198k both times: 1st: Processing: Node(428597k) Way(32778k) Relation(198k) Reading in file: - Entity: line 3: parser error : Document is empty

Re: [OSM-dev] This week's planet .osm

2009-09-13 Thread Thomas Wood
It sounds like the file you have is incomplete, check the md5 hash against http://planet.openstreetmap.org/planet-090909.osm.bz2.md5 2009/9/13 Richard Ive rich...@xanox.net: Hey all, Has anyone had any issue with this weeks planet dump? 090909 I've tried to rum an osm2pgsql import twice now

Re: [OSM-dev] This week's planet .osm

2009-09-13 Thread Richard Ive
Thanks, but they match: File: dfcbe4f2e98b7e0a10dd9470a031a439 planet-090909.osm.bz2 [postg...@maps osm2pgsql]$ md5sum ../planet-latest.osm.bz2 dfcbe4f2e98b7e0a10dd9470a031a439 ../planet-latest.osm.bz2 2009/9/13 Thomas Wood grand.edgemas...@gmail.com It sounds like the file you have is

Re: [OSM-dev] This week's planet .osm

2009-09-13 Thread Martin Lesser
Richard Ive rich...@xanox.net writes: Has anyone had any issue with this weeks planet dump? 090909 ... Has anyone else had an issue at all? Not sure whether my issues are the same as yours but I had at minimum two problems - with relation 226076 (containing a tab in member type=node

[OSM-dev] Postgres 8.4/PostGIS 1.4 performance

2009-09-13 Thread Frederik Ramm
Hi, I did a full (osm2pgsql) planet import on a standard Ubuntu Jaunty system and it took 1794 minutes. Then I upgraded to Postgres 8.4 (backported from Karmic) plus PostGIS 1.4 (home-built package), and re-tried the import: 2028 minutes (that's 13% performance loss). Can anybody confirm

Re: [OSM-dev] Postgres 8.4/PostGIS 1.4 performance

2009-09-13 Thread Jon Burgess
On Sun, 2009-09-13 at 19:56 +0200, Frederik Ramm wrote: Hi, I did a full (osm2pgsql) planet import on a standard Ubuntu Jaunty system and it took 1794 minutes. Then I upgraded to Postgres 8.4 (backported from Karmic) plus PostGIS 1.4 (home-built package), and re-tried the import:

Re: [OSM-dev] This week's planet .osm

2009-09-13 Thread Martin Lesser
Ignore this please - it's not the reason, osm2pgsql handles both correctly. The encoding error at my site was caused by something else. Martin Lesser ml-osm-...@bettercom.de writes: Not sure whether my issues are the same as yours but I had at minimum two problems - with relation 226076

[OSM-dev] Taxi to Lisbon

2009-09-13 Thread Claudius Henrichs
I've set up a wiki page [1] with some impressive routing results on OSM data compared to Google. For example we can already route from Teheran, Iran to Lisbon, Portugal and to New Delhi, India. Or you can hail a cab in Mexico City and tell the driver to take you to Anchorage, Alaska. Feel free

Re: [OSM-dev] Postgres 8.4/PostGIS 1.4 performance

2009-09-13 Thread Frederik Ramm
Hi, Jon Burgess wrote: Try running the query below and compare the sizes returned for the tables indexes on the two databases. Hm, it seems I have overlooked the fact that I also updated osm2pgsql when I did the update. The old database was created with 0.66. But it already had integers in

Re: [OSM-dev] Taxi to Lisbon

2009-09-13 Thread jamesmikedup...@googlemail.com
Wow, this is great. thank you for the great software! mike On Sun, Sep 13, 2009 at 10:37 PM, Claudius Henrichs claudiu...@gmx.de wrote: I've set up a wiki page [1] with some impressive routing results on OSM data compared to Google. For example we can already route from Teheran, Iran to

Re: [OSM-dev] Postgres 8.4/PostGIS 1.4 performance

2009-09-13 Thread Jon Burgess
On Sun, 2009-09-13 at 22:45 +0200, Frederik Ramm wrote: Hi, Jon Burgess wrote: Try running the query below and compare the sizes returned for the tables indexes on the two databases. Hm, it seems I have overlooked the fact that I also updated osm2pgsql when I did the update. The old

[OSM-dev] Whitespace in tag-keys

2009-09-13 Thread Martin Lesser
I did not find explicit informations about this issue in the wiki so I'm wondering whether whitespace in keys is a should not or a must not. What's the recommended action if one detects tags like eg. http://www.openstreetmap.org/browse/node/275671906 (k=is in)

Re: [OSM-dev] Whitespace in tag-keys

2009-09-13 Thread Dermot McNally
2009/9/13 Martin Lesser ml-osm-...@bettercom.de: I did not find explicit informations about this issue in the wiki so I'm wondering whether whitespace in keys is a should not or a must not. I think you're in fact addressing two separate issues here. As we can see, the API will allow you to

Re: [OSM-dev] Whitespace in tag-keys

2009-09-13 Thread Thomas Wood
2009/9/13 Martin Lesser ml-osm-...@bettercom.de: I did not find explicit informations about this issue in the wiki so I'm wondering whether whitespace in keys is a should not or a must not. What's the recommended action if one detects tags like eg.  

Re: [OSM-dev] Postgres 8.4/PostGIS 1.4 performance

2009-09-13 Thread Frederik Ramm
Hi, I think I made one more blunder when doing the comparison: I did one of the imports - the slower one actually! - with the -l flag, the other one without. (Both were with --slim.) But by now there are too many variables in the game and it is impossible to tell what is to blame for the

Re: [OSM-dev] This week's planet .osm

2009-09-13 Thread Jon Burgess
On Sun, 2009-09-13 at 16:37 +0100, Richard Ive wrote: Hey all, Has anyone had any issue with this weeks planet dump? 090909 I've tried to rum an osm2pgsql import twice now and it's falled over at Relation 198k both times: Relation 198k is the end of the file. It sounds like you have

Re: [OSM-dev] Whitespace in tag-keys

2009-09-13 Thread Cartinus
On Sunday 13 September 2009 23:28:09 Martin Lesser wrote: A second question about ways/relations with less than two nodes/members: Is there any reason for these? A relation with only one member could be a stub of something that will have more members once completely mapped. -- m.v.g.,

Re: [OSM-dev] Postgres 8.4/PostGIS 1.4 performance

2009-09-13 Thread Martin Lesser
Frederik Ramm frede...@remote.org writes: Jon Burgess wrote: Do you have the _int.sql loaded? Yes, in both versions; however the _int that comes with PostGIS 1.4 seems to be slightly modified, Has postgis its own _int.sql? AFAIK _int.sql is part of postgres-contrib, in the postgis src I

Re: [josm-dev] [PATCH 1/2] QuadBuckets (using quad tiling) for node storage in JOSM

2009-09-13 Thread Karl Guggisberg
Hi Dave Any idea how you want this to look? I'm starting to code some primitive things, and it would be nice if I could get whatever I do merged eventually. AFAIK, nothing has been done so far. * Object updates are rare, and it is OK if the notification process is relatively slow * Rather

[josm-dev] JOSM Version check

2009-09-13 Thread Dirk Stöcker
Hello, The version check for JOSM now moved totally to the server. Instead of [[Version/running)]] which added a comment later replaced by JOSM now a macro [[VersionTest(1|2|3|4)]] exists, which has 4 texts as argument. 1) Text for normal version display 2) Text for outdated version 3) Text

[josm-dev] Filters (something as long awaited layers) for JOSM

2009-09-13 Thread Petr Dlouhý
Hello, I have made patch adding filters dialog to OSM http://josm.openstreetmap.de/ticket/3475. Please test it (I will try to send compiled josm jar in next email), and send me your opinion on that. -- Petr Dlouhý ___ josm-dev mailing list

[josm-dev] Coders: Please use true quotation marks

2009-09-13 Thread Claudius
I'm currently translating JOSM and stumble every now and then over strings like Value ''{0}'' is going to be applied for key ''{1}'' I understand that it's easier not needing to escape quotation marks, but replacing them with two single quotes is plain wrong. So I'm addressing all JOSM

Re: [josm-dev] Coders: Please use true quotation marks

2009-09-13 Thread Karl Guggisberg
Hi Claudius Value ''{0}'' is going to be applied for key ''{1}'' It's not as wrong as it looks. The strings are interpreted by the tr(...) function which treats '' as ', whereas as sinlge ' disappears. Users will see one ' only. Whenever you see '' in I18n string, keep both in the translation or

Re: [josm-dev] Coders: Please use true quotation marks

2009-09-13 Thread Matthew W. S. Bell
On Sun, 2009-09-13 at 19:46 +0200, Claudius wrote: Thanks from the translators (and lovers of correct punctuation), Actually, I believe that the use of quotes, italics, etc. to distinguish between the use of a phrase and the mention of a phrase[1] is a style rule only. Matthew Bell [1]

[josm-dev] Table syntax in MOTD

2009-09-13 Thread Ævar Arnfjörð Bjarmason
I made a MOTD which uses HTML table syntax: http://u.nix.is/~avar/motd.html I think it uses up space a bit better than the current one. But unfortunately trac deosn't support HTML tables (they end up being escaped). 1. http://josm.openstreetmap.de/wiki/StartupPageSource