download area display (yellow stripes) rectangular also for non-mercator projection

[skyper]

Hi

It is happening for around 10 days now, that not all data is downloaded at the boarder of the "downloaded area" or maybe the area is rendered to big. Anyway I end up with node within the downloaded area where not all connected ways are shown. This is dangerous. It does not happen all the time and it is not that easy to reproduce ( I tried it but failed so far.)

It happened several times last week !

Maybe it has something to do with lambert-9 projection, but I am not sure about it.

cu skyper

r3652

[skyper]

I happened again. I downloaded some data with merkator-projection and then switch to lambert-CC9. The downloaded area is rendered to big.

[anonymous]

I tried but could not see problem. If you can give a step by step description how to reproduce the problem, then it will be much easier for us to fix it. The difference can be as large as "impossible to fix" vs. "fixed in 1 minute". :)

[bastiK]

anonymous is me

[skyper]

Replying to anonymous:

I tried but could not see problem. If you can give a step by step description how to reproduce the problem, then it will be much easier for us to fix it. The difference can be as large as "impossible to fix" vs. "fixed in 1 minute". :)

At least I can show you what happens. I attached a osm-file. It was downloaded with mercator-projection, but if you switch to Lambert-cc9 (7) you notice that the downloaded area is rendered bigger and objects which where outside the downloaded area with mercator are now within it.
Have a look at the buildings in the lower right corner.

r3688

[skyper]

example file

[stoecker]

The "downloaded area" is drawn rectangular, which is no longer true for non-mercator projection. I'm not sure if the added processing load required to fix this issue is really a good idea.

[bastiK]

Then we'd also have to draw curved way segments. When you draw a way segment from west to east, the center of the line can be off by a few cm or m in north south direction, depending on the length of the segment. I found the effect to be negligible with Mercator projection and normal sized way segments (<1km).

That is, if a way segment is defined by "shortest line on the sphere" and not by "shortest line in Mercator". :)