Desktop: System.Security.Cryptography.Rsa.Tests.KeyGeneration.GenerateKey_4096 failed with "Xunit.Sdk.EqualException"

[KristinXie1]

Failed test; System.Security.Cryptography.Rsa.Tests.KeyGeneration.GenerateKey_4096

Detail: https://ci.dot.net/job/dotnet_corefx/job/master/job/outerloop_netfx_windows_nt_debug/29/testReport/System.Security.Cryptography.Rsa.Tests/KeyGeneration/GenerateKey_4096/

MESSAGE:

Assert.Equal() Failure
    Expected: 4096
    Actual: 1024

STACK TRACE:

at System.Security.Cryptography.Rsa.Tests.KeyGeneration.GenerateKey(Func`2 getSize) 
in D:\j\workspace\outerloop_net---903ddde6\src\Common\tests\System\Security\Cryptography\AlgorithmImplementations\RSA\KeyGeneration.cs:line 57 
at System.Security.Cryptography.Rsa.Tests.KeyGeneration.GenerateKey(Int32 size) 
in D:\j\workspace\outerloop_net---903ddde6\src\Common\tests\System\Security\Cryptography\AlgorithmImplementations\RSA\KeyGeneration.cs:line 43 
at System.Security.Cryptography.Rsa.Tests.KeyGeneration.GenerateKey_4096() 
in D:\j\workspace\outerloop_net---903ddde6\src\Common\tests\System\Security\Cryptography\AlgorithmImplementations\RSA\KeyGeneration.cs:line 38

[KristinXie1]

This issue is repro on full framework tests, detail: https://mc.dot.net/#/product/netcore/master/source/official~2Fcorefx~2Fmaster~2F/type/test~2Ffunctional~2Fdesktop~2Fcli~2F/build/20170425.01/workItem/System.Security.Cryptography.Algorithms.Tests/analysis/xunit/System.Security.Cryptography.Rsa.Tests.KeyGeneration~2FGenerateKey_2048

[krwq]

netfx tiny repro:

            using (RSA rsa = RSA.Create()) // CSP impl
            {
                rsa.KeySize = 4096;
                Console.WriteLine(rsa.KeySize); // =>1024
            } 

This is clearly incorrect. @bartonjs, is it ok to disable this for now on netfx and port this change later? This is clearly not .NET Core related and needs some change ported.

[krwq]

Same bug: https://github.com/dotnet/corefx/issues/12388

[krwq]

Ok this is apparently by design according to that - I'm disabling those on netfx

[bartonjs]

Yep, I've finally come over to @morganbr's way of thinking and think we should fix RSACryptoServiceProvider; probably with a retargeting quirk. It's pretty interesting watching how far one way or the other I go down the feature vs compat arguments in my head.