More lib testing with native AOT

[LakshanF]

Experimental PR to get test coverage for native aot with some libraries

[dotnet-issue-labeler]

I couldn't figure out the best area label to add to this PR. If you have write-permissions please help me learn by adding exactly one area label.

[MichalStrehovsky]

Note this didn't trigger any of the NativeAOT test legs. We definitely want to run them before merging this.

[MichalStrehovsky]

The FileSystem test was failing with an OOM on Linux.

I've added a commit that disables pre-enumaration of theories. It's what the "official" runner also does:

https://github.com/dotnet/arcade/blob/7ffde50ec98dbe3584bef2375165ec9a6b39cb35/src/Microsoft.DotNet.XUnitConsoleRunner/src/ConsoleRunner.cs#L338-L341

The FileSystem tests have several tests that allocate big byte arrays as TheoryData that we then have to keep in memory. My theory is that disabling the pre-enumeration relieves the memory pressure:

runtime/src/libraries/System.IO.FileSystem/tests/File/Copy.cs

Lines 93 to 108 in d201ebb

	public static IEnumerable<object[]> CopyFileWithData_MemberData()
	{
	var rand = new Random();
	foreach (bool readOnly in new[] { true, false })
	{
	foreach (int length in new[] { 0, 1, 3, 4096, 1024 * 80, 1024 * 1024 * 10 })
	{
	char[] data = new char[length];
	for (int i = 0; i < data.Length; i++)
	{
	data[i] = (char)rand.Next(0, 256);
	}
	yield return new object[] { data, readOnly};
	}
	}
	}

[MichalStrehovsky]

Looks good.

@LakshanF the Linux ARM64 Helix workitem actually failed a test. Could you look into why the CI lane still succeeded?

https://helixre107v0xdeko0k025g8.blob.core.windows.net/dotnet-runtime-refs-pull-69132-merge-1662767c930a41fbb3/System.Threading.Tests/1/console.b615871b.log?helixlogtype=result

I got to this log by clicking through the "Send to Helix" step on https://dev.azure.com/dnceng/public/_build/results?buildId=1801875&view=logs&jobId=b2c675c1-1c32-50e9-7c40-e762459300db&j=b2c675c1-1c32-50e9-7c40-e762459300db&t=4f51db95-f7de-5a4e-cd10-f078782dd75b

Following the JSON at "Sent Helix Job; see work items at https://helix.dot.net/api/jobs/1662767c-930a-41fb-b34d-b2c8e1f71572/workitems?api-version=2019-06-17" and following more JSON from there to get to the console output (not sure if there's a more convenient way to get to that).

Making sure CI lanes with failing tests don't succeed is a Pri-0.

Cc @dotnet/ilc-contrib FYI.

[MichalStrehovsky]

The libraries test run failure is #70128, merging.

[LakshanF]

I got to this log by clicking through the "Send to Helix" step on https://dev.azure.com/dnceng/public/_build/results?buildId=1801875&view=logs&jobId=b2c675c1-1c32-50e9-7c40-e762459300db&j=b2c675c1-1c32-50e9-7c40-e762459300db&t=4f51db95-f7de-5a4e-cd10-f078782dd75b

Strange, I no longer can access the above link to see the failing job but can see the log results with the 1 failure.

I can follow the now successful job in Linux-arm64 since you disabled the test via "https://helix.dot.net/api/jobs/1662767c-930a-41fb-b34d-b2c8e1f71572/workitems/System.Threading.Tests?api-version=2019-06-17"

Thanks a lot for all the investigation and the work to merge this PR!

[MichalStrehovsky]

Strange, I no longer can access the above link to see the failing job but can see the log results with the 1 failure.

that's odd. It was this run: https://github.com/dotnet/runtime/runs/6701362762. I now get an error when trying to follow the link to pipelines.

Either way, we now have the same problem in the System.Reflection test mentioned on Teams. That test is failing in a lot of the open PRs, but the CI lane is green.