Fixed renting buffer from pool in INumberBase (dotnet#102755)

Fixed xml-doc comments in OptionsServiceCollectionExtensions.AddOptio…

…nsWithValidateOnStart (dotnet#94385)

Fixed wrong comment in Marshalling/*Marshaller.cs (dotnet#92563)

For the BufferSize 0x200 the maximum size is 512 (dec).

Handle the remainder in MemoryExtensions.Count vectorized (dotnet#82687)

* Re-use targetVector from Vector256 in Vector128 code-path

Saves an additional vpbroadcastb.

* Process remainder vectorized

* Use unsigned division for remaining elements

Avoids the signed integer division.

* Process remainder vectorized only if more than half of vector remains

Benchmarking showed that the cost is quite high, so for just a few elements the scalar loop seems better.

* Added comment why / 2 got chosen

Optimizations for Ascii.Equals and Ascii.EqualsIgnoreCase (dotnet#85926)

Base64.Decode: fixed latent bug for invalid input that is less than a…

… block-size (dotnet#79952)

* Tests

* Fix

Fixed relative links in Memory-model.md (dotnet#79785)

MemoryExtensions.Replace(Span<T>, T, T) implemented (dotnet#76337)

* Defined API

* Tests

* Scalar implementation

* Use EqualityComparer<T>.Default instead

* Delegation to SpanHelpers.Replace

* ReplaceValueType implemented

* Use ushort instead of short, as it doesn't sign-extend for broadcast and in the scalar loop

* Forward string.Replace(char, char) to SpanHelpers.ReplaceValueType

* Process remainder vectorized only when not done already and with max width available

* Split into inlineable scalar path and non-inlineable vectorized path

* Replaced open coded loops with Replace

* Don't use EqualityComparer<T>.Default

Cf. dotnet#76337 (comment)

* Remove guards for remainder

Cf. dotnet#76337 (comment)

* Don't split method into scalar and vectorized and don't force inlining of scalar-part

* Fixed assert

ReplaceValueType is called from string.Replace(char, char) so the Debug.Assert was on wrong position, as at entry to method non accelerated platforms are allowed to call it.

* Better handling of remainder from the vectorized loop(s)

Intentionally leave one iteration off, as the remaining elements are done vectorized anyway. This eliminates the less probable case (cf. dotnet#76337 (comment)) that the last vector is done twice.

* PR feedback

[release/7.0] Base64.Decode: fixed latent bug for non-ASCII inputs (d…

…otnet#76812)

* Added test cases

* Fix

* Use unicode escape sequence in tests

* Apply suggestions from code review

Co-authored-by: Günther Foidl <[email protected]>
Co-authored-by: Adam Sitnik <[email protected]>

Base64.Decode: fixed latent bug for non-ASCII inputs (dotnet#76795)