Simplify scaleline calculation

[ahocevar]

It looks like the scaleline control has some legacy code that is no longer needed now that the projection provides a point resolution in crs units.

Fixes #4572.
Fixes #3444.

[flipper44]

This is a picture of a one square mile block and a resulting metric scaleline, after zooming the scaleline crashes. The map is using a us-ft projection. All layers Bing ESRI etc are reprojecting fine.
I do not get a crash when using metric projection.
I will take a look at where it is breaking

[ahocevar]

Could you please provide more context @flipper44, i.e. the code or a JSFiddle you used to produce the above image?

[flipper44]

Its hard to show the Ful;l code because it is dynamically built.
Here is what I found when using this Projection WKT
WKT ="PROJCS["IN83-EF",GEOGCS["LL83",DATUM["NAD83",SPHEROID["GRS1980",6378137.000,298.25722210]],PRIMEM["Greenwich",0],UNIT["Degree",0.017453292519943295]],PROJECTION["Transverse_Mercator"],PARAMETER["false_easting",328083.333],PARAMETER["false_northing",820208.333],PARAMETER["scale_factor",0.999966666667],PARAMETER["central_meridian",-85.66666666666670],PARAMETER["latitude_of_origin",37.50000000000000],UNIT["Foot_US",0.30480060960122]]";
proj4.defs('Custom', Wkt);
var x = proj4.defs('Custom');
var newProj = ol.proj.get('Custom');

Everything works up until my first Zoom (scale Bar values are still wrong) because viewState.Resolution is NaN PS this also causes ZoomSlider to crash

if I set newProj.units_ = "us-ft";
then it works (still bad scale values)
Edit:
I have found that I must do this because the value for that comes from proj4js and OL is "foot_us" and not "us-ft"..This WKT came from a Mapguide Runtimemap

[flipper44]

It looks like I will have to manually handle units coming from mapguide for now.

The Fix for bad Scale Values:
I Havent tested it fully but it would seem this line should be changed from
projection.getPointResolution(viewState.resolution, center) /
metersPerUnit;
to
projection.getPointResolution(viewState.resolution, center) *
metersPerUnit;

[ahocevar]

You're right @flipper44, the point resolution needs to be multiplied by the meters per unit. I made that change. Now I'll create an example with your WKT def to see if I can reproduce what else is wrong.

[ahocevar]

@flipper44 513cf13 fixes your other issue. I also added an example.

This is now ready for review.

[tschaub]

My computer stops working when I get to 5. Might be worth limiting this to 6.

[tschaub]

Could also be that enforcing a center constraint with extent would help.

[tschaub]

Looks like there is an issue when setting units: 'degrees'.

Really nice to see the removal of the unnecessary code here.

[flipper44]

Re-projection on smaller local coordinate systems does get odd when zOomed way past the local extents.
Also I think there was a reason I had an if statement for degrees in my initial example

[ahocevar]

Zoom constraint changed, center constraint introduced, degree issue fixed.

[fredj]

LGTM