#4111 closed defect (fixed)
editing certain ways at certain points makes them temporarily unselect-/editable
| Reported by: | malenki | Owned by: | team |
|---|---|---|---|
| Priority: | major | Milestone: | |
| Component: | Core | Version: | latest |
| Keywords: | Cc: | malenki, florian@… |
Description
At some places ways are not selectable/editable anymore after dragging a part of them. After undoing the edit vthe way remains in this state. To crossing ways happens the same. After selecting some other way editing is again possible.
Reproducibility is difficult, here a detailed manual:
Download data in Josm using this bounding box and drag at the straight part of the curve where the marker is placed. To see the marker, oben the link in a browser
Josm version 2585, appeared also in several versions before
Attachments (1)
Change History (16)
comment:1 follow-up: 2 Changed 10 years ago by
comment:2 Changed 10 years ago by
Replying to jttt:
Does it happen everytime when you follow the instructions?
Yes. Confirmed by othr users
I have trouble reproducing it.
Maybe you drag at the wrong part of the way? The affected part at the example is quite small. Perhaps zooming close to the marker helps?
When you select other way, everything is back to normal? Meaning you can also select the modified way?
After selecting some other way I am able to select again the previous unselectable way, yes.
comment:3 Changed 10 years ago by
| Cc: | malenki added |
|---|---|
| Priority: | normal → major |
Since in the last days I got this behaviour quite often but unfortunately not as reproducible as the above one it's in my eyes not a funny singular thing anymore.
comment:4 follow-up: 5 Changed 10 years ago by
PS: when this happens now the ways stay unselect/-editable so the workflow is seriously affected.
comment:5 Changed 10 years ago by
| Cc: | florian@… added |
|---|
Replying to malenki:
PS: when this happens now the ways stay unselect/-editable so the workflow is seriously affected.
I also got this bug but also not as reproducible. After restart JOSM i can't reprouduce the error. From time to time this error occures on a arbitrary way. Sometimes i think there's a random seed in JOSM.
comment:7 Changed 10 years ago by
When this happen again, you can try to run Dataset consistency test (in Osm data layer popup menu).
About #4324 - if the way disappears after repaint (you can move map a bit to cause repaint) then it might be also bug in QuadBuckets. But I think these too tickets are not related.
comment:8 follow-up: 9 Changed 10 years ago by
Does is still happen with the original bbox? The marker is at the middle of segment. Does it mean you create new virtual node by dragging the cross or move the segment?
comment:9 Changed 10 years ago by
Replying to jttt:
Does is still happen with the original bbox?
yes
The marker is at the middle of segment. Does it mean you create new virtual node by dragging the cross
yes
I will add a screenshot to point which way is meant. It shouldn't matter (anymore?) at which virtual node (cross) of it you pull.
Changed 10 years ago by
| Attachment: | gkrellShoot_2010-01-17_josm_drag_for_disabled_way.png added |
|---|
comment:11 follow-up: 12 Changed 10 years ago by
Replying to jttt:
And dataset consistency test found any errors?
Don't know what this is or how to perform. Please explain. (in IRC nobody answered, in wiki nothing is found)
comment:12 follow-up: 13 Changed 10 years ago by
Replying to malenki:
Replying to jttt:
And dataset consistency test found any errors?
Don't know what this is or how to perform. Please explain. (in IRC nobody answered, in wiki nothing is found)
Right click on data layer in layer list dialog, second item from bottom in the menu is "Dataset consistency test". It will run some test and show list of errors if anything is found.
In screenshot I see you use wireframe mode instead of mappaint. It means this ticket might be related to QuadBuckets after all. You can try it with r2868, maybe it got fixed.
comment:13 follow-up: 14 Changed 10 years ago by
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
Replying to jttt:
Right click on data layer in layer list dialog, second item from bottom in the menu is "Dataset consistency test". It will run some test and show list of errors if anything is found.
thx
In screenshot I see you use wireframe mode instead of mappaint. It means this ticket might be related to QuadBuckets after all.
Why should josm behave different depending on the display mode?
You can try it with r2868, maybe it got fixed.
With r2869 this error doesn't appear anymore. (I just pulled the latest version from svn.)
comment:14 Changed 10 years ago by
Replying to malenki:
Why should josm behave different depending on the display mode?
Mappaint use QuadBuckets to paint only primitives that are currently visible. Wireframe is fast enough even without this so it doesn't use this optimalization yet.
comment:15 Changed 10 years ago by
Well, I should probably use QuadBuckets for wireframe mode as well. I just did some tests and it's currently way slower than mappaint for big datasets.
Does it happen everytime when you follow the instructions? I have trouble reproducing it. When you select other way, everything is back to normal? Meaning you can also select the modified way?