ENH: allow start=range_object in RangeIndex constructor

[jbrockmendel]

• closes API: should the RangeIndex constructor accept start=range_index? #12067
• tests added / passed
• Ensure all linting tests pass, see here for how to run them
• whatsnew entry

simplifies Index.__new__ a bit

[jreback]

hah I originally implemented this but then we settled on .from_range. I would be +1 in using this. cc @jorisvandenbossche

[topper-123]

I'm +1 on this change. If we call something RangeIndex, it's reasonable that it takes a range in its constructor IMO.

[jorisvandenbossche]

(no strong opinion here)

[jreback]

lgtm. I think ok to also deprecated .from_range

[topper-123]

In the RangeIndex class doc stringm in the Parameters section, I think the line start : int (default: 0), or other RangeIndex instance should be changed to start : int (default: 0), range or other RangeIndex instance

[jbrockmendel]

In the RangeIndex class doc stringm in the Parameters section, I think the line start : int (default: 0), or other RangeIndex instance should be changed to start : int (default: 0), range or other RangeIndex instance

good idea. will add this to my upcoming core.indexes TYP branch if thats OK

[jreback]

thanks @jbrockmendel

followups when you can