@Konstantin-Posudevskiy in .NET I get overflow exception, but not in python. Also precision in .NET is much less for some reason:

Python 2.7.13 |Anaconda custom (32-bit)| (default, Dec 19 2016, 13:36:02) [MSC v
.1500 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
Anaconda is brought to you by Continuum Analytics.
Please check out: http://continuum.io/thanks and https://anaconda.org
>>> dmax = 1.7976931348623159e308
>>> dmax
inf
>>> dmax = 1.7976931348623156e308
>>> dmax
1.7976931348623155e+308
>>> dmax = 1.7976931348623154e308
>>> dmax
1.7976931348623153e+308
>>> dmax = 1.7976931348624159e308
>>> dmax
inf
>>> dmax+1e308
inf

scriptcs (ctrl-c to exit or :help for help)

> double dmax = 1.7976931348623159e308d
* ;
(2,15): error CS0594: Floating-point constant is outside the range of type 'doub
le'
> double dmax = 1.7976931348623156e308d;
> dmax
1.79769313486232E+308
> dmax = 1.7976931348623146e308d;
1.79769313486231E+308

@denfromufa, I absolutely agree with you.

This pull request addresses another issues. It only

1. fixes invalid usages of Python C API method 'PyFloat_AsDouble';
2. supports conversion from Python 'inf' and '-inf' to PositiveInfinity and NegativeInfinity for System.Double (using method 'PyFloat_AsDouble')
3. supports conversion from Python float value to PositiveInfinity and NegativeInfinity for System.Single. As I am concerned, Python float values, which are in range of 4 bytes can be represented with the similar byte representation of System.Single in .NET, even if they are values of different precision.
   OverflowException is thrown for values, that are represented with 5-8 bytes
4. adds tests for paragraphs (2) and (3).

I also checked failing Python tests. Please, take a look at method test_double_conversion.
Python tests expects for .NET overflow exception, in case when it passes 'inf' value. IMO, it is incorrect, as infinity is valid double value.

Python output for test values from test method test_double_conversion:

>>> dmax = 1.7976931348623159e308
>>> dmax
inf
>>> dmin = -1.7976931348623159e308
>>> dmin
-inf

@Konstantin-Posudevskiy can you fix this failing test?

        with pytest.raises(OverflowError):
>           ConversionTest().DoubleField = 1.7976931348623159e308
E           Failed: DID NOT RAISE <type 'exceptions.OverflowError'>

OverflowError is not .NET exception, it is Python exception and it should not be expected in this case.

@denfromufa sure, I will do.

Codecov Report

Merging #487 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #487      +/-   ##
==========================================
+ Coverage   74.23%   74.24%   +<.01%     
==========================================
  Files          64       64              
  Lines        5562     5564       +2     
  Branches      894      894              
==========================================
+ Hits         4129     4131       +2     
  Misses       1147     1147              
  Partials      286      286

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update bfce5e2...315d2d1. Read the comment docs.