Re: [OSM-dev] Duplicated nodes in Potlatch
Am Sonntag 05 Oktober 2008 13:01:48 schrieb Richard Fairhurst: Marc Schütz wrote: Several of the ways in this area http://www.openstreetmap.org/? lat=49.89487lon=10.88733zoom=17layers=0B00FTFTTT have been modified with Potlatch so that they contain the same node twice in a row. Is this a known bug? Should the API be changed to reject such ways? It's not expressly forbidden to have a way comprising 865234 occurrences of the same node in succession, though I agree it's not desirable, so strictly it's not a bug. Potlatch does already trap most such occurrences (see http:// trac.openstreetmap.org/browser/applications/editors/potlatch/ way.as#L691) but if you can provide steps to reproduce that would be helpful - on a trac ticket, please, so I can keep track of it, and ideally with way ids as I've just spent two minutes clicking around and can't actually find which ways you're talking about. ;) Sorry, I thought I had added the IDs. I will contact the user and ask him what exactly he did. Regards, Marc signature.asc Description: This is a digitally signed message part. ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
Marc Schütz wrote: Several of the ways in this area http://www.openstreetmap.org/? lat=49.89487lon=10.88733zoom=17layers=0B00FTFTTT have been modified with Potlatch so that they contain the same node twice in a row. Is this a known bug? Should the API be changed to reject such ways? It's not expressly forbidden to have a way comprising 865234 occurrences of the same node in succession, though I agree it's not desirable, so strictly it's not a bug. Potlatch does already trap most such occurrences (see http:// trac.openstreetmap.org/browser/applications/editors/potlatch/ way.as#L691) but if you can provide steps to reproduce that would be helpful - on a trac ticket, please, so I can keep track of it, and ideally with way ids as I've just spent two minutes clicking around and can't actually find which ways you're talking about. ;) cheers Richard ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
Rogier Wolff wrote: Number one has the disadvantage of allowing the database to grow bigger than it needs to be. Heh. Of course, we have an elephant in the room there - one which, mercifully, 0.6 will address. I committed a change last night (not deployed yet) so that amf_controller filters out created_by tags on nodes before sending them to Potlatch, which ignores them anyway. On average it cuts the amount of data transmitted by about 25% cheers Richard ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
[OSM-dev] Duplicated nodes in Potlatch
Several of the ways in this area http://www.openstreetmap.org/?lat=49.89487lon=10.88733zoom=17layers=0B00FTFTTT have been modified with Potlatch so that they contain the same node twice in a row. Is this a known bug? Should the API be changed to reject such ways? Regards, Marc signature.asc Description: This is a digitally signed message part. ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
On Sun, Oct 5, 2008 at 1:02 PM, Richard Fairhurst [EMAIL PROTECTED] wrote: Rogier Wolff wrote: Number one has the disadvantage of allowing the database to grow bigger than it needs to be. Heh. Of course, we have an elephant in the room there - one which, mercifully, 0.6 will address. I committed a change last night (not deployed yet) so that amf_controller filters out created_by tags on nodes before sending them to Potlatch, which ignores them anyway. On average it cuts the amount of data transmitted by about 25% just out of interest, i did a quick check on a recent UK extract: 65% of all tags are created_by and that 87% of created_by tags are JOSM. i was quite surprised :-) cheers, matt ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
Matt Amos wrote: just out of interest, i did a quick check on a recent UK extract: 65% of all tags are created_by and that 87% of created_by tags are JOSM. 'taint a simple comparison, though: Potlatch doesn't set created_by on nodes at all, but does set created_by on way updates. JOSM sets created_by on way and node creation, but not on update. I did a few bits of SQL footling around recently on a UK planet and it would _roughly_ suggest that more UK edits at present are done with Potlatch than JOSM. I suspect the reverse is true in Germany! cheers Richard ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
On 05/10/2008 17:00, Matt Amos wrote: On Sun, Oct 5, 2008 at 1:02 PM, Richard Fairhurst [EMAIL PROTECTED] wrote: Rogier Wolff wrote: Number one has the disadvantage of allowing the database to grow bigger than it needs to be. Heh. Of course, we have an elephant in the room there - one which, mercifully, 0.6 will address. I committed a change last night (not deployed yet) so that amf_controller filters out created_by tags on nodes before sending them to Potlatch, which ignores them anyway. On average it cuts the amount of data transmitted by about 25% just out of interest, i did a quick check on a recent UK extract: 65% of all tags are created_by and that 87% of created_by tags are JOSM. i was quite surprised :-) And this underestimates it because Potlatch replaces any created_by's on objects originally created by JOSM, but JOSM doesn't replace Potlatch's created_by's. You can see the same effect in ItoWorld's OSMMapper: often Potlatch is used to tweak, but JOSM used for the heavyweight bulk mapping. David ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
2008/10/5 David Earl [EMAIL PROTECTED]: And this underestimates it because Potlatch replaces any created_by's on objects originally created by JOSM, but JOSM doesn't replace Potlatch's created_by's. The underestimation is probably worse than that as JOSM stoped creating created_by's at all a while ago. I'm not at all familiar with the JOSM codebase but it looks like the changes in OsmServerWriter.java here are the cause. http://josm.openstreetmap.de/changeset/633 Now only the changeset gets a created_by tag but since api 0.6 isn't hear yet I guess they just get ignored. So anybody using josm-latest that is less than 5 months old won't be adding created_by tags at all. -- DavidD ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
David Earl wrote: And this underestimates it because Potlatch replaces any created_by's on objects originally created by JOSM, but JOSM doesn't replace Potlatch's created_by's. Which is _really_ annoying for debugging purposes. ;) But anyway, you are, as an unwise man once said, misunderestimating this. Matt was talking about objects, and like I say, Potlatch doesn't write created_by on nodes (as of 0.10d, it won't even _read_ them). There are currently 300 million nodes vs 27 million ways. cheers Richard ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
Am Sonntag, den 05.10.2008, 17:10 +0100 schrieb Richard Fairhurst: I did a few bits of SQL footling around recently on a UK planet and it would _roughly_ suggest that more UK edits at present are done with Potlatch than JOSM. I suspect the reverse is true in Germany! See also http://www.mail-archive.com/dev@openstreetmap.org/msg03175.html Sincerely, Joachim ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
Hi, this thread is really valuable for me as someone who writes an osm editor. With respect to overwriting the created_by entry, i was just copying potlatchs behaviour. So: What's the preferred behaviour? Overwriting imho doesn't make much sense. But since osm2go is in an early development change i thought it makes much sense to be able to leave fingerprint of my tool, so that in case of major trouble caused by a stupid bug in my tool there's a chance to find where my tool has done harm. Why is this something the application does, anyway? The issue is mho very similar to the user assignment which is entirely done on server side. Since the clients all identify tjhemselves in their upload why doesn't the server also handle this? The created_by is like the user a technical thing and not really information related to the physical objects described. So i never really understood why this is a tag and not an entirely different thing like the user is. Till Am Sonntag 05 Oktober 2008 schrieb Richard Fairhurst: David Earl wrote: And this underestimates it because Potlatch replaces any created_by's on objects originally created by JOSM, but JOSM doesn't replace Potlatch's created_by's. Which is _really_ annoying for debugging purposes. ;) But anyway, you are, as an unwise man once said, misunderestimating this. Matt was talking about objects, and like I say, Potlatch doesn't write created_by on nodes (as of 0.10d, it won't even _read_ them). There are currently 300 million nodes vs 27 million ways. cheers Richard ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
i was just copying potlatchs behaviour. So: What's the preferred behaviour? Merkaartor systematically overwrite the created_by when doing changes on the server (after having followed a similar thread). If this tag has any value, it is, IMHO, to be able to identify editors having strange/wrong behaviors needing to be addressed. AFAIK, there is no standard way for the server to know which editor is sending it data. It could, based upon the user-agent, but I don't think there something foreseen regarding this. - Chris - ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Duplicated nodes in Potlatch
The idea is to move the created_by tag to the changeset for the API0.6, thus reducing the amount of data stored considerably. It would also mean that any changeset could only be used for one editor. Shaun On 5 Oct 2008, at 21:34, Till Harbaum / Lists wrote: Hi, this thread is really valuable for me as someone who writes an osm editor. With respect to overwriting the created_by entry, i was just copying potlatchs behaviour. So: What's the preferred behaviour? Overwriting imho doesn't make much sense. But since osm2go is in an early development change i thought it makes much sense to be able to leave fingerprint of my tool, so that in case of major trouble caused by a stupid bug in my tool there's a chance to find where my tool has done harm. Why is this something the application does, anyway? The issue is mho very similar to the user assignment which is entirely done on server side. Since the clients all identify tjhemselves in their upload why doesn't the server also handle this? The created_by is like the user a technical thing and not really information related to the physical objects described. So i never really understood why this is a tag and not an entirely different thing like the user is. Till Am Sonntag 05 Oktober 2008 schrieb Richard Fairhurst: David Earl wrote: And this underestimates it because Potlatch replaces any created_by's on objects originally created by JOSM, but JOSM doesn't replace Potlatch's created_by's. Which is _really_ annoying for debugging purposes. ;) But anyway, you are, as an unwise man once said, misunderestimating this. Matt was talking about objects, and like I say, Potlatch doesn't write created_by on nodes (as of 0.10d, it won't even _read_ them). There are currently 300 million nodes vs 27 million ways. cheers Richard ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] way 27483626 UTF-8 truncation
On Sun, Oct 5, 2008 at 2:12 AM, Matt Amos [EMAIL PROTECTED] wrote: On Sat, Oct 4, 2008 at 9:36 AM, Florian Lohoff [EMAIL PROTECTED] wrote: To get the ROMA database in sync again i replaced the notes by broken-utf8 - As notes typically get not rendered thats not a problem for me though. ROMA was down for a half a day before i discovered the broken files and fixed them ... likewise. the easiest way to fix them was hand-editing the change files. i don't find it to be particularly onerous - just the price we pay for being on the bleeding edge ;-) The corrupted data in the db has been fixed by TomH and I've re-generated the changeset files. Unfortunately it's too late for you guys now ... ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [josm-dev] Commit access
On Sat, 4 Oct 2008, Henry Loenwind wrote: BTW: SVN holds 11 patches, now. I applied some stuff and also #1622. Thought I'm not yet happy with it. a) Please save keyboard configuration only if it differs from the defaults to allow later changes. b) Don't save automatic changes. These should be visible in shell to be fixable. c) The list is not sorted in any way I find useful. d) The entries should use the same translation, as the menus. e) I see elements in group edit, whereas they are in the tools menu. f) I would reduce the list to: 1) menu (when applicable) 2) action (translated one, but otherwise unmodified) 3) shortcut Everything else should be internal. For the next time I consider JOSM to be in unstable mode and don't recommend latest installations except for testing and bug fixes. Ciao -- http://www.dstoecker.eu/ (PGP key available) ___ josm-dev mailing list josm-dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/josm-dev
[josm-dev] Information for plugin authors about new keyboard configuration
Dear plugin authors, when you'll recompile your pligins against the latest JOSM you'll get a number of deprecated warnings. You might be tempted to change your code to use the new API, and usually I would ask you to do so. However, I just noticed that I forgot to include a way to revoke the automatically assigned shortcuts. So if you use the new API and use correct IDs instead of generated ones, your original shortcut will be unavailable to you. Bad thing. Also, using the new API means that the changed plugin will no longer work with older JOSM version. So please wait with changing your plugins. However, I noticed that at least one of the existing plugins (Surveyor) uses the old API incorrectly. Please make sure to use the shortcut property of JosmAction to register your shortcut with Swing, not you own KeyStroke object. Thank you Henry ___ josm-dev mailing list josm-dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/josm-dev