I had a chance to drive past there today and stopped at parking lot for survey. Well, actually there's no signs about being a parking lot. Put together quick drawing of possible osm features. https://cdn.discordapp.com/attachments/537315188039221301/892523230919786496/parnamae-narva.png
In essence, what you see on Mapillary (and streetview) is up to date and valid. No-entry sign comes after the parking lot entrance, but the public part does go across two lowered kerbs (thin white line on drawing).
Most crossings have standardized tactile paving (drawn as thick white line) except crossing on Linnuse tee, which uses mosaic paving (green). Linnuse tee can be entered from all 3 roads and turning to all 3.
Notable odd thing is crossing in parking lot (near tip of yellow arrow) which has traffic_calming=table only on parking lot's side, but not on Linnuse tee's side. SW cycleway has guard rail along northern side, but there's gap at southern corner of parking lot.
All 4 directions of the junction have traffic lights. While U-turn at Irusilla tn is possible, it's not realistic due to volume of traffic and narrow road section.

I think Linnuse tee (way/129226648) should be mapped as residential road, but since it's dead-end road, there's no point in adding complicated access restriction tags.

Current state of Linnuse tee is tagged as cycleway with
motor_vehicle=destination on segment from parking aisle in north until concrete block in south. This means that routing engines will direct drivers to the residents via parking lot and sometimes via Irusilla. Luckily, traffic signs on this junction (#map=19/59.46023/24.90521) are visible on about month old Mapillary imagery.

* Turning right from Narva mnt to Linnuse tee is allowed as per destination sign on https://www.mapillary.com/app/?pKey=1001641713992040
* So is left-turn also allowed: https://www.mapillary.com/app/?pKey=2870427889890013
* Notice how first sign is black-on-white while second is white-on-blue. That is caused by junction being located on the legal border of city.
* Traffic signs (No entry except residents) on Linnuse tee are visible on https://www.mapillary.com/app/?pKey=132240762253013
. * Going straight from Pärnamäe tee to Linnuse tee is also allowed as per that image series.
. * There are no visible signs on Linnuse tee and median on Narva mnt has dashed line. Therefore You can turn on all 3 directions from Linnuse tee. It has traffic lights btw.
. * The living-street sign is placed roughly at the spot where street moves from one jurisdiction to another, BUT it's rotated 90 degrees compared to what it should be. That is technique often used with temporary road construction signs to indicate that restrictions are not applied at the moment.
* As for southern end of the street, there are metal barriers visible on last week imagery, but i can't determine, if there are more traffic signs. https://fotoladu.maaamet.ee/etak.php?B=59.45880109804291&L=24.90564843766783&fotoladu
* Regarding conflicting traffic signs, both Tallinn, Jõelähtme and maybe Viimsi should be notified, because road follows along border of two.

No idea if caused by this changeset, but node/633025611 may need fixing as well. I am not sure if it has tactile paving though. Looks like non-standard rumble strip.

PS. This junction was recently redrawn on OSM and anyone fixing this may encounter inconsistent relations or lanewise tagging.

Tere, Kas on mingi oluline põhjus, miks on vaja osad sõidurajad nii pikkade eraldiseisvate joontena joonistada? (rõhk sõnal "pikkade")
Näiteks pööre Pärnamäe teelt Narva mnt-le võiks olla ~6-7 korda lühem (350 vs 50 m; way/979841381). Ma saan täiesti aru eri suundadeks jagamise kasulikkusest, aga mitte sellest, miks nad nii vara peavad hargnema. Ametlikult öeldakse vist, et tee peaks hargnema füüsilise takistuse juurest (ohutussaar keset teed) või erandjuhul pidevjoone algusest, aga mitte kunagi uue pöörderaja algusest.
Tegelikult näevad ka mõned bussiliinid selle muudatuse tulemusel kummalised välja (vt Iru bussipeatus), aga nende pärast pole hetkel vaja muretseda.

I'm really sorry. Situation was caused because i used overpass query for only (high)ways and nodes, but didn't include relations. Issue is impacting up to 1219 ways in 13 changesets in 2 cities.
Such changes were made in Tallinn, Tartu and Haapsalu. Latter was done with iD, so no relations were broken there.

Should these changes be reverted (and later in far future properly recreated) or is it worth attempting to fix up? In first case, are there solutions to split ways based on ends of current ways while properly slicing relations as well? In latter case useful tools might be Josm's relations' plugin and OSM inspector (http://tools.geofabrik.de/osmi/?view=pubtrans_routes). I have neither competence to fix the situation nor time to readd lanes (using iD) in case of revert. These splittings didn't affect only bus routes, but also road route relations as well.

Affected changesets:
changeset/109705088
changeset/109706438
changeset/109706513
changeset/109708590
changeset/109710326
changeset/109710842
changeset/110061689
changeset/110064342
changeset/110076297
changeset/110383287
changeset/110409629
changeset/110423518
changeset/110424431

PS. Thanks for your incredible effort you have previously shown to add those bus routes to OSM!

Ma hakkasin kirjutama kommentaari, et Maa-ameti kaart ei näitagi poodide asukohti, aga tuleb välja, et mõned üksikud ettevõtted on kaardile siiski kantud [1]. Samuti kuvatakse Terviseameti kaartidel apteeke [2] ja ujulaid [3]. See, et kasutaja poolt allikana kasutatav andmekogu (nt Maa-ameti kaart) ei sisalda teatud nähtuste klassi (nt alkoholipoode), ei tähenda, et nimetatud nähtuseid ei eksisteeriks.

[1] - https://xgis.maaamet.ee/xgis2/page/link/oJW2jm0y
[2] - https://xgis.maaamet.ee/xgis2/page/link/I4N9yY02
[3] - https://xgis.maaamet.ee/xgis2/page/app/terviseamet_veetervis

Objektide liigutamine pika vahemaa taha ongi muudetud võimalikult keeruliseks, sest tihti on juhtunud, et keegi võtab kinni mõnest ristmikust ning tõstab selle kogemata paralleeltänavale. Näiteks saaks Gonsiori tänav teha läbi Kadrioru pargi hüppe Narva maanteele.

Viga parandatud kogumis changeset/109665545

Tere,
Tekkis küsimus, kas Truumu diagnostika töökoda asub Petrooleumi tänaval või Pärnu maanteel? Hetkel jääb käesolevast muudatusest mulje, nagu oleks ettevõte avanud Petrooleumi tn-l teise esinduse ning Pärnu mnt-l lammutatud majas asuv esindus on endiselt samas kohas avatud, kuid on vahepeal saanud uue aadressi. Kas tegu on veaga ning kas võin duplikaadi eemaldada, või soovid ise proovida olukorda parandada? Juhul kui ettevõte kolis ühest kohast teise, on mõistlik võtta vana (Pärnu mnt-l asuv) variant ja tõsta see Petrooleumi tn-le. Hetkel samas kohas asuva sõlme võib kustutada.
Ette tänades, Fghj753.

Ok then. Thanks for replying.

Hello,
I noticed you are using custom script to upload bus stops to OSM. Due to slightly diagonal shape Norway, changesets covering entire Norway like this one will also cover most populous parts of neighbouring countries, which may upset some people. Is it possible to split your changesets somewhere between latitude 64..66N?
I took a look at your upload script and it's processing based on counties. Seems like easiest solution to implement would be running two copies of script: one which excludes counties ["18", "54"]; other version excludes others ["03", "11", "15", "30", "34", "38", "42", "46", "50"]. Downside of this approach means having to input your password twice. Alternative is perhaps segregating lines 974-994 to separate function, passing excluded counties to function via arguments and calling it twice.
Do you mind looking into splitting changesets?

Parandatud. Review no longer needed.

Thanks for route_master relations tip.
I think relations not being edited in this set was probably due to limitation of Level0 editor, which only loads first 500 elements. Relations are usually listed at the end of overpass results set and therefore editor dropped them due to meeting limit on maximum number of elements.

Sorry for global changeset. I tried to introduce bounding box to overpass query but global query was only way to run it without getting any errors.
Changed elements where value of colour=* was clearly hexadeciamal number (matched regex `^[0-9a-fA-F]{6}$`), but didn't have prefixed `#`.
Query was [out:xml][timeout:60]; ( nwr["colour"~"^[0-9a-fA-F]{6}$"];); out meta;

@qqqqqqqqqqqqqqqqqqqq
I tried to craft together overpass query to find deleted elements, but it didn't work reliably as it didn't find control sample way/61732730 (https://pewu.github.io/osm-history/#/way/61732730). Similar situation is also at #map=19/59.40476/27.28695 where's 3 overlapping parks and apartment building overlapping a road.
While there are many Pokemon Go related issues in the world of OSM, their in-game maps haven't been updated for years and I would assume most players don't see point in updating OSM as their changes won't show up in game. Also this user's actions involve more than just parks, for example apartment buildings and footpaths such as "пешеходная дорожка в лесу 2" (way/467983228) north of town. Pay attention to access=no tags on these paths.
As for the real Rahvapark, looks like it haven't been added before. Closest candidate would be way/926197944. Btw, i think only reason Rahvapargi tiik has preserved it's name was because of it's wikidata tag, as iD doesn't let users to easily rename objects with wikidata connection.
Please have a look at avenue aside Kohtla-Järve park (https://pewu.github.io/osm-history/#/way/48237896). I find it unlikely, that main street of town is two parallel pedestrian streets with sidewalks. Furthermore, there's Maxima grocery store's parking lot on that street with entrance from park's side and on aerial imagery from 30th of May (https://fotoladu.maaamet.ee/?basemap=kiirortofoto&zlevel=15,27.28616,59.40216) cars are still allowed on the street. I tried to fix that road few days after it was edited and my change was virtually reverted within day.

I have been occasionally observer their actions and sometimes written down notes when something is notably odd. "Park for recreation number 4" (way/938285913) for example is part of cluster of 4 intersecting parks, but that particular park has address "пешеходная дорожка 66" named after nearby footway. Officially, all that forest should be called one big Metsapargi.
Just like Keskallee, there are more legit-named streets which have been converted to footways: way/61732621, way/60542348, way/60542353, way/234531443.

ENG:
Comments of this changeset and changeset/106293325 suggest adding new street, but both changesets add just two overlapping parks: парк тихий (way/954028662) and парк отдыха (way/954028810) (quiet park
and recreation park respectively).
While I doubt about validity of two overlapping parks and their names, I think it's safe to conclude that these parks are not streets.
Please use comments that reflect reality.

EST:
Käesoleva ning järgneva muudatuskogumi kommentaarid väidavad, nagu oleks muudatuste raames lisatud kaardile uued tänavad, samas kui tegelikkuses lisasite kaks teineteisega kattuvat parki: парк тихий (way/954028662) and парк отдыха (way/954028810), mille nimed tähendavad tõlkes vastavalt vaikne park ja puhkepark.
Kuigi ma kahtlen nende kahe pargi nimedes ning kattuvuses, olen kindel, et käesoleval juhul on need ikkagi pargid, mitte tänavad.
Palun kasutage selliseid kommentaare, mis kirjeldavad tegelikkust.

RUS (g-translate):
Комментарии к этому пакету правоки и changeset/106293325 предлагают добавление новую улицу, но оба пакети изменений добавляют только два перекрывающихся парка: "Парк тихий" (way/954028662) и "Парк отдыха" (way/954028810).
Хотя я сомневаюсь в достоверности двух перекрывающихся парков и их названий, я думаю, что можно с уверенностью заключить, что эти парки, не улицы.
Пожалуйста, используйте комментарии, которые отражают реальность

PS: Bing is the worst imagery in Estonia. Please use Esri or Maa-amet layers.
Bing худший аэрофотоснимки в эстонии. Используйте аэрофотоснимки от Esri или Maa-amet.

Tutorials/guides/Учебники:
osm.wiki/RU:Changeset
osm.wiki/RU:Хороший_стиль
osm.wiki/Armchair_mapping
osm.wiki/Verifiability

Hello, Thanks for stepping by.
You seem to be experienced russian-speaking OSM mapper. Parks you deleted were added by relatively new russian-speaking user tetrab0rat, who has added many suspiciously named and overlapping parks all over the town; one park per changeset. Sadly I haven't contacted them directly because i don't speak russian.
Do you mind leaving some constructive criticism to one of tetrab0rat's recent changesets?

PS. Please consider using more informative changeset comments in the future, such as "deleted bunch of parks".

Hello. Please refrain from using Bing imagery in Estonia as it's possibly the worst imagery available here. Please use either Esri or Maa-amet imagery.
In your use case it might be good idea to familiarize yourself with Maa-amet's pre-release imagery: https://fotoladu.maaamet.ee/?basemap=kiirortofoto&zlevel=14,24.70670,59.45319&overlay=avaleht

You can select background imagery with shortcut key "B".

PS. It might be good idea to use more informative chageset comments rather than merely single letter h.

@josh_joplin_lp Turned out tests failed because specific command used to run tests was changed on Wednesday. I couldn't try out Vietnam on my computer, because after 2 hours it had filled entire my hard drive, but it worked fine with Cambodia (one of smallest map files for Asia).
I have updated the scripts and despite all the errors DB test gives, most of them were about integration tests covering signing in with third party accounts. There were also some errors about user diary entries and moderation functionalities, but they seemed to work.
I was able to create some changes in Josm and upload them to server.

PS. If you wish to add iD to your website as well, you need to follow some easy steps at https://github.com/openstreetmap/openstreetmap-website/blob/master/CONFIGURE.md#oauth-consumer-keys .

@josh_joplin_lp. Sorry for late reply. Looked up on my VM, that uses just Estonia. Database size was 9GB, while pbf file was 83MB. I think you will need about 27-28GB for database, but then you'll also need additional space for all possible files that may be created during your process. For example I noticed my OSM server had 3 GB of logs.
Considering Ubuntu OS itself needs ~15 GB, I'd suggest to have around 50GB virtual disk for VM. Osmosis runtime for me was around 30-40 min. Since your file 3 times bigger, it should take 1.5-2 hours (highly variable depending on your hardware).
Optionally, take a look at your host PC's power settings and make sure that it's power mode is set to high performance or automatic. Mine went had around 40% faster after changing power mode.

@josh_joplin_lp Use install_english.sh and on line 7 and 8 change pbf_filename to what you like, pbf_download_url to suitable region's download url. I think you may want "https://download.geofabrik.de/asia/vietnam-latest.osm.pbf".
On lines 124, 132 and 133 you can define your dev account user's username and password. Ln 126, 127 can be changed to whatever you want, but you don't have to.

After installing, use run.sh to start api server.

Hi,
Interesting question. I know that editors usually assign negative numbers as temporary IDs for elements, but that's also where my knowledge on the subject end.
It's not allowed to use main database for development purposes, but you should either use dev server (https://master.apis.dev.openstreetmap.org) or set up one yourself.
I personally had similar problem few months ago and i created Bash script to set up copy of OSM-API on local Ubuntu 20.04 VM. https://github.com/kallejre/geo/tree/main/building%20geometry%20importer/API%20server%20setup I hope it helps.
You can configure JOSM to use different OSM server other than default main API.