* Clean up code emitted for single-character loops

We can emit a simple while loop rather than a while true loop with gotos and (now unnecessary) labels.

* Clean up inverted loop in FindFirstChar

* Clean up minimum length check to be more idiomatic

* Clean up MatchCharacterClass condition to be more idiomatic

* Address PR feedback

Microsoft.NET.ILLink.Tasks
 From Version 7.0.100-1.21479.1 -> To Version 7.0.100-1.21501.1
Co-authored-by: Ndotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>

This is a better fix for #58373 that changes the handling of this to
happen in morph for all cases. We sometimes missed the insertion of
necessary casts because we forgot to remove a GTF_DONT_CSE flag when
folding an indirection. Fixing this leads to some new GT_CNS_DBL cases
in lowering that hit an assert, but those cases should be correctly
handled by the default case so just remove the assert.

To get rid of some of the regressions I have allowed generating
assertions when assigning struct fields from casts. It was unclear why
this was not allowed in the first place.

* Move runtime rolling builds to run twice a day on main

* Update eng/pipelines/runtime.yml

The move was pretty straightforward and I didn't really need to touch anything.

The only changes I made:
* crossgen2command.txt is now command.txt
* crossgen2repro.rsp is now repro.rsp
* Instead of "CrossGen2 version" the TXT file now says "Compiler version". What compiler will be obvious from the file name and version strings.

* Make libraries tests use the same config as the matching libraries build.

This change is useful in two ways:

1. It simplifies our CI build dependency graph so that we don't have a windowx x86 job depending on a windows x64 job to finish first.
2. It unblocks #59579 which introduces an arch-dependent native component into the libraries tests with DNNE.

* Mark one of the win x86 runs as release.

* Remove testBuildPlatforms.

* Build debug Linux_arm/Linux_musl_arm/Linux_musl_arm64 libraries builds on PR.

* Update to only new libraries Debug builds to run on a coreclr change in PR.

* Upload test artifacts with os subgroup name

* Use a "preprocessor if" instead of a condition for isFullMatrix.

* Fix conditions based on feedback.

* Fix incorrect condition.

* Don't run tests locally when we're running tests using Helix.

* Use ProductVersion in how-to-debug-dump

* Refactoring XslCompiledTransform

* remove unused methods

* review feedback
Co-authored-by: Nkronic <kronic@softland.ru>

* Remove 3 active issues

* Apply suggestions from code review
Co-authored-by: NStephen Toub <stoub@microsoft.com>
Co-authored-by: NStephen Toub <stoub@microsoft.com>

* Do not flush serializer buffers before first IAsyncEnumerable element is fetched

* add debug assertion

* address feedback

Co-authored-by: Ndotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>

Co-authored-by: Ndotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>

[main] Update dependencies from dotnet/arcade dotnet/hotreload-utils dnceng/internal/dotnet-optimization (#59256)

[main] Update dependencies from dotnet/arcade dotnet/hotreload-utils dnceng/internal/dotnet-optimization


 - Merge branch 'main' into darc-main-61786f4e-24d3-4a12-aded-829376797d8c

Conflicts:
	eng/Version.Details.xml
	eng/Versions.props

* Fix an assert and tighter check for arm64 misestimation

* Align methods at 32B

`ILGenerator.Emit(short)` does not use short forms if applicable; only the int overload does, but here we can directly use `ldarg.0`.

The RtlRestoreContext sets SP before reading out PC from the
context. That can lead to a corruption of the PC in the context
if an async signal is delivered to the thread  or the thread is
interrupted by any other mean after the SP is set and before
the value of PC is extracted from the context.

This change fixes it by setting the SP after both PC and SP values
are read from the context data structure.

OSR methods must do special setups for register parameters, loading them
from the original method's frame. Update the computation we do here so it
works for TYP_STRUCT register params.

Closes #59875.

* Refactoring Vector<T> to be alphabetically ordered and reduce duplicated code

* Adding a debug view to System.Numerics.Vector<T>

* Do some basic code cleanup and refactoring of Vector64/128/256

* Updating Vector64/128/256 to expose cross platform helper methods

* Adding templated tests for Vector64

* Regenerating templated tests for Vector64

* Adding templated tests for Vector128

* Regenerating templated tests for Vector128

* Adding templated tests for Vector256

* Regenerating templated tests for Vector256

* Fix getSIMDStructFromField to account for accessing the private ulong fields for Vector64/128/256<T>

* Ensure helper intrinsics don't insert on importation

* Minor cleanup of impBaseIntrinsic for x86/x64

* Intrinsify the Vector64/128/256 methods

* Make the internal helper named IsTypeSupported to not conflict with mono's existing check

* Ensure we lie about the type for TYP_SIMD32 bitwise ops when only AVX is supported

* Use gtNewSimdZeroNode rather than gtNewSIMDVectorZero

* Applying formatting patch

* Apply suggestions from code review
Co-authored-by: NSingleAccretion <62474226+SingleAccretion@users.noreply.github.com>

* Apply suggestions from code review

* Split HardwareIntrinsic tests into 3 groups

* Update src/coreclr/vm/class.cpp
Co-authored-by: NEgor Chesakov <Egor.Chesakov@microsoft.com>
Co-authored-by: NSingleAccretion <62474226+SingleAccretion@users.noreply.github.com>
Co-authored-by: NEgor Chesakov <Egor.Chesakov@microsoft.com>

If there's a P/Invoke to a library, direct `netcore_probe_for_module` to look
in NATIVE_DLL_SEARCH_DIRECTORIES, followed by the directory containing the
managed assembly.

Fixes https://github.com/dotnet/runtime/issues/59766

- Use `ConfigurationSources.Add`
- Update the comment

* Obsolete Regex.CompileToAssembly

It's never worked on .NET Core and for .NET 7 we'll now have the regex source generator.

* Remove argument validation from CompileToAssembly

Per PR feedback

* Split and rework fgMorphCast

Split it into the part that does pre-order transforms
and the part that does post-order optimizations.

Rename the pre-order part to "fgMorphExpandCast" and let
"fgMorphSmpOp" do the morphing of operands and optimization.

Just one diff for this commit: "gtFoldExprConst" does not
retain the types of handles, thus LSRA was not seeing
"long" and "byref" constant integers with the same values as
equivalent and did not reuse the register.

* Delete unneeded code from "fgOptimizeCast"

These cases are now covered by "fgMorphSmpOp".

* Refactor comma throw propagation

So that it works well for casts.

There are diffs for this commit in the ILGEN methods,
mostly positive, but some are regressions.

The regressions fall into 2 buckets:
  1. The new code creates overflow helper blocks for casts
     before propagating comma throws, thus leaving them
     around, dead. This is a problem in general and should
     be solved by moving the helper insertion later.
  2. The new code doesn't propagate comma throws after global
     morph for casts. This is is by design, as the throw
     propagation does not work well with VNs (and CSE).

The improvements are due to new code setting "fgRemoveRestOfBlock"
for comma throws originating from casts.

Invoke-WebRequest throws an exception when the download fails, so we
should use try-catch instead of status code to check failure. We can use
the generic Retry function from tools.ps1 to do this.

In addition pass -PassThru to avoid leaving a corrupted download file in
case the download fails. This will buffer the download and write it once
at the end.

Fix #57196

* Fix comparison in EmitMultiChar

For strings of length 2 we weren't using the optimized path of reading/comparing a single uint.

* Remove duplicative length checks for some concatenations

When we have a sequence like `abc[def]ghi`, today we'll emit a length check of 3 for the `abc` multi, a length check of 1 for the `[def]` set, and a length check of 3 for the `ghi`, but we can instead emit a single length check for 7.

This also fixes a small bug where I used a > instead of a >= in a length comparison, the net effect of which was using two comparisons instead of getting away with one for a two character string.

And added some debug-only output that helped me diagnosing why things were working the way they were.

* Special-case regexes where Go can be a nop

For an expression like "abc", if FindFirstChar returns true, it's found and fully validated the match.  There's no benefit to comparing it again in Go; instead, Go can simply perform the capture.

* Clean up generated Boyer-Moore code

* Fix NRT warnings in source generated code

* address feedback

* address style

* Increase timeout for ConsoleLifetimeExitTests.EnsureEnvironmentExitDoesntHang

* missed the comment

* fixes to get Quic working on macOS

* Apply suggestions from code review
Co-authored-by: NMarie Píchová <11718369+ManickaP@users.noreply.github.com>
Co-authored-by: NMarie Píchová <11718369+ManickaP@users.noreply.github.com>

For better or worse, Regex matches char by char, without any attention paid to surrogate pairs and the like.  In .NET Core 3.0, a change was made to the regex parser that used ToLower on a whole string rather than calling ToLower on each individual char, but this then leads to inconsistencies with how the matching is performed, making it so that things that used to match no longer do.

* Add CWDE support to the emitter and codegen

It is a shorter form of a sign extending mov from eax to eax.
Use it in codegen for casts.

* Fix printing for CWDE and CDQ

Their size determines what they do.

* Move the optimization to emit_Mov

This allows us to catch more cases.