* Unifying many of the R2R and COR constants between crossgen2 and the r2rdump tools

* Move tools/crossgen2/Common to tools/Common

* Adjust paths in csproj files

* Initialize compSwitchedToOptimized and compSwitchedToMinOpts before calling to compInitOptions in jit/compiler.cpp

* Add LF to switchedToOptimized and switchedToMinOpts messages in jit/compiler.cpp

* Fix getting affinity set on MUSL on Jetson TX2

The code in PAL_GetCurrentThreadAffinitySet relied on the fact that the
number of processors reported as configured in the system is always
larger than the maximum CPU index. However, it turns out that it is not
true on some devices / distros. The Jetson TX2 reports CPUs 0, 3, 4 and
5 in the affinity mask and the 1 and 2 are never reported. GLIBC reports
6 as the number of configured CPUs, however MUSL reports just 4. The
PAL_GetCurrentThreadAffinitySet was using the number of CPUs reported as
configured as the upper bound for scanning affinity set, so on Jetson
TX2, the affinity mask returned had just two bits set while there were
4 CPUs. That triggered an assert in the GCToOSInterface::Initialize.

This change fixes that by looping over all cpu indices in the affinity set.
Similar fix went to GetProcessorForHeap and related stuff in gcenv.unix.cpp

These are left-over from .NET Native. Not needed for RyuJIT or Mono.

Fix crossgen2 handling of direct calls to abstract methods

It seems that the InvalidProgramException is more appropriate.

This change fixes two differences between old and new crossgen in the
getCallInfo method. Direct calls to abstract methods were being compiled
as if they were possible and lead to runtime crash in the call chain
from the PreStubWorker.

Fix missing unbox flag on method handle for struct methods

* Delete common upload artifact template

* Move upload artifact step from coreclr to common

We force `/OPT:REF` and this is incompatible with incremental linking
(and yields a warning to that effect). To fix, explicitly disable
incremental linking.

(I couldn't find anyplace where it was explicitly enabled, so I believe
it is enabled by default with the given linker options.)

Another option would be to not force `/OPT:REF` for Checked builds.

* Adding some tests for marshalling generics.

* Updating the VM to allow marshalling blittable generic types.

* Adding comments for why certain blittable generics are blocked from being marshaled

* Removing the new IDS_EE_BADMARSHAL_BLITTABLE_GENERICS_RESTRICTION string in favor of fixing the text in IDS_EE_BADMARSHAL_GENERICS_RESTRICTION

* Updating CrossGen2 to handle marshalling blittable generics

* Add a script for generating a .sln containing all library projects

* Add a script for generating a .sln containing all library projects

The compiler is updating its handling of maybe-null values, and in doing so we're getting a bunch of new warnings.  We've not yet ingested the new compiler, but this proactively addresses the warnings for when we do.

It is just the TPA binder context now.

Remove dead code

* Move more shared components to common and move helix queues out of platform-matrix

* PR Feedback

* Add Libraries Testing framework package as inline

The existing package Microsoft.DotNet.CoreFxTesting lived in Arcade
because of no infrastructure being available to compile local tasks in
the repository. As the runtime repository now offers that we can inline
the testing framework.

* Hardcode configuration for installer.tasks

* Update ReportGenerator global tool version

* Add vstest support

* Update binary serialization blobs

Fixes Loader\classloader\generics\regressions\dev10_393447 test runtime
failure when compiled with crossgen2.

This mimicks what old crossgen does.

Fix devirtualization in crossgen2

* propogate assertions through Exception handlers.

* Add comments.

* Fix undocumented generatelayout command

* Address feedback

Port more Pinvoke fixes from the single-exe branch

This is only for arm64 musl to address #170.

* Remove some indirection when raising AssemblyResolve event
* Delete unused code

Note this is an unforutunate loss of coverage.
It ups the priority on trait based tests or fixing the root cause.

Enable warn as error in corelib as the workaround was set for VS2017 which we no longer use since we moved to VS2019 and that's the recommended toolset.

* Update links in docs/libraries

* Fix more docs/libraries links

CoreCLR subrepo build has a known existing warning coming from
pgomgr that Arcade upgrades to an error via WarnAsError. I propose
suppressing autodetection of warnings in CoreCLR build to fix this
spurious bug reproducing in release mode.

Thanks

Tomas

Fixes: #150
Follow-up tracking issue: #156 

Turns out zero length arrays are passed as a pointer to the non-existing element only when the array is passed by reference.