NullPointerException when uploading changes

[leo]

happens in versoin 1566
i made a few changes to an area in austria (2h work)
when uploading the changes, i get an exception (

[Moloc]

I also get a NullPointerException when uploading with version 1566. (JOSM 1529 reports IOException: 412 PreConditionFalied)

Maybe this is useful to find the error?

[cyron@…]

confirming.

NullPointerException when uploading with ref 1566.

Adding wireshark log and OSM which fail for me.

[cyron@…]

wireshark log and osm which fails - part1

[cyron@…]

wireshark log and osm which fails - part2

[cyron@…]

wireshark log and osm which fails - part3

[avarab@…]

I am unable to reproduce this, on 1570 I get "Version mismatch: Provided 2, server had: 3 of Node 488050". That's because after trying to upload this the reporter ​made another changeset in the area, so the bug might still be reproducable if that hadn't been the case.

[leopold]

I just re-tried it with version 1569 (latest), and it still has the same error.
on the upload-bar there is written: ~"Finishing Changeset" (translated from german)
i will attach my osm file.

[leopold]

part 1/4

[leopold]

part 2/4

[leopold]

part 3/4

[leopold]

part 4/4

[frawe]

At least Molocs and Cyrons are caused by the same reply as ticket 2509.

According to Tom Hughes the API should send content with the error but doesn't (see ​http://trac.openstreetmap.org/ticket/1783). JOSM nevertheless should catch the error better than NPE and at least show "Precondition failed".

[stoecker]

Ticket #2509 has been marked as a duplicate of this ticket.

[anonymous]

Ticket #2493 has been marked as a duplicate of this ticket.

[jeff@…]

I'm having the same problem, except on 1569 (1578 and 1580 crash on me for other reasons). I'll attach the file I'm working with.

[jeff@…]

Well, I would have attached the file, except it says that it's spam. It's just a small osm file bzipped to get under the file size limits.

[anonymous]

When you create an account and login your bonus points are higher and you can upload the file.

[jcollie]

osm file that gives null pointer exception when trying to upload

[leo]

for my case the problem is solved now.

i just downloaded Version 1583, and tried it again. I also got the "Prediction failed" error, but now with more details of the error:
it said something like "Node XXX is still used in way XXX"
do not know how this could happen, but I opened the .osm in my notepad++, and deleted the action="deleted" tag from the node that was named in the error message.
after this, the upload worked :)

[frawe]

Replying to leo:

for my case the problem is solved now.

It sort of is because the server now delivers content in case of 412 which it didn't before (see ticket 1783 I mentioned before).
I think it this tickets priority could be lowered to minor but it should still get fixed in JOSM.

[anonymous]

I think it this tickets priority could be lowered to minor but it should still get fixed in JOSM.

Think so too. JOSM now reads both error messages replied from the server in error headers and in error documents. But how this information is processed and presented to the user should still get fixed, see also #2437. For instance, JOSM could automatically do what leo did manually using notepad++. And users could be directed to a help page with more verbose (and translated) information about the error message.

[leo]

showing the Error Message is great. also fixing this automatically would be great.
But: why i think this error should not occure, as i edited the map with JOSM.

I mean the follwoing: it seems that i deleted a node N that belongs to a way W.
for me, it seems that JOSM should have updated the list of nodes, that belongs to way W - and this is the real bug, isn't it?

but by the way: thanks for you great help and work to JOSM!!

[Gubaer]

I mean the follwoing: it seems that i deleted a node N that belongs to a way W.
for me, it seems that JOSM should have updated the list of nodes, that belongs to way W - and this is the real bug, isn't it?

I can imagine scenarios where a user deletes a node, JOSM correctly updates its local datasets (sets the node to deleted, removes the node from all ways and relations in its local dataset), but the server still replies a 412 Precondition Failed.

What if somebody else created a new way including the node and uploaded his changes before you could upload your changes? Only the server could detect this when you upload your changes. And it would probably reply a 412 Precondition Failed.

JOSM today doesn't support users well when the server detects conflicts of this kind. But from what I could observer so far it correctly manipulates its local dataset.

[stoecker]

Closed as duplicate of #2609.