You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
All VSTest cares about is having the runtimeconfig json file, AFAIK. For that, it uses the ugly trick of setting OutputType to Exe. This solution has been for many years, but it doesn't seem like it was a good idea. I think a better idea might be to set GenerateRuntimeConfigurationFiles to true so that the runtimeconfig.json is generated.
The text was updated successfully, but these errors were encountered:
One note is whether or not we delete the entry-point generation. In case the user is adding OutputType Exe in csproj, it will be breaking if we didn't generate the entrypoint. I'm not sure how common that scenario is though. But we may just keep the entrypoint generation under Condition="'$(OutputType)'=='Exe'".
Another note is interaction with MTP. Maybe some users of MSTest metapackage and MTP are relying on VSTest to set the OutputType for them. But I think it may be fine to break those as it's not a good idea, and eventually causes some other issues in certain scenarios where the .NET SDK has some logic relying on whether or not OutputType is Exe, but that logic of .NET SDK may be evaluated before VSTest is able to set OutputType. So it was already fragile.
All VSTest cares about is having the runtimeconfig json file, AFAIK. For that, it uses the ugly trick of setting OutputType to Exe. This solution has been for many years, but it doesn't seem like it was a good idea. I think a better idea might be to set GenerateRuntimeConfigurationFiles to true so that the runtimeconfig.json is generated.
The text was updated successfully, but these errors were encountered: