Moving code that is not valid in WRC from Common code to NetFramework… #460
Conversation
| /// <summary> | ||
| /// Get nullable $quantityName from nullable $($unit.PluralName). | ||
| /// </summary> | ||
| public static $($quantityName)? From$($unit.PluralName)($($quantityValueType)? $valueParamName) |
There was a problem hiding this comment.
Outside this PR, but in light of binary size issues - nullable From methods add twice the method count for all our 800 units and I'm contemplating removing them from the lib in vNext. Thoughts?
It's only a convenience thing the way I see it, and this pattern is also a kind of weird pattern I've not seen anywhere else in .NET or third party libs. I would instead advocate checking null explicitly or using Option<T> or similar to box null value handling.
There was a problem hiding this comment.
Generally I think that the consuming library would store the nullable type.
For example they would have an Acceleration? field. Then they either assign it somewhere, or not. I would generally not expect a null to come into a ?FromXXX() method. Rather the unit would just never be constructed.
I'm all for removing them.
More work towards #409. I pulled out all code hidden behind the following preprocessor directive and moved it to the NetFramework code generator.