Add net461 to targetFrameworksList - #1961
Conversation
safern
left a comment
safern
left a comment
There was a problem hiding this comment.
ML.NET is using net461 to run their tests. So they need a lower version of assets in order to resolve to the right set of dlls from xunit extensions dlls for full framework.
@ViktorHofer can we remove net472 and just cross compile it for net461? I guess there wouldn't cause any issues since the source code is the same.
Sure we can't downgrade the .NET Framework version to net461, that shouldn't cause any issue.
I would prefer upgrading their version to net472. Isn't that feasible? |
did you mean can here ? @ViktorHofer is there any specific reason for choosing the net472 ? |
|
Yeah can sorry. Choosing the highest possible framework version gives us more flexibility but if ML.NET can't support net472 I'm fine with downgrading it to the highest supported one. |
|
I agree with @ViktorHofer that is best to be in the latest framework. Is there any reason why ML.NET can target net472? |
|
We can close this one as ml people are moving towards not using conditionalFact entirely dotnet/machinelearning#2402 |
3 participants
Currently when we use this package in netfx version like 461 , the nuget resolves this version to netstandard.
The conditionalFact attributes need to be cross compiled which is not possible in netstandard , hence the conditionalFact always evaluated the condition to be false.