Avoid unnecessary array allocations in ZipArchive

[stephentoub]

And simplify a few open-coded loops into span operations.

[dotnet-policy-service]

Tagging subscribers to this area: @karelz, @dotnet/area-system-io-compression
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

Reduce GC pressure and simplify inner-loop code paths in System.IO.Compression by reusing buffers from ArrayPool<byte> and replacing manual loops with span operations.

Changes:

• Replace allocating range operations / intermediate arrays with pooled buffers when reading the Zip central directory (sync + async).
• Fix an allocation-heavy path in ZipCentralDirectoryFileHeader.TryReadBlockAsync by collating into a pooled Memory<byte> rather than creating a new array.
• Simplify repeated writes in InflaterManaged.DecodeDynamicBlockHeader using Span.Fill / Span.Clear.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.

File	Description
src/libraries/System.IO.Compression/src/System/IO/Compression/ZipBlocks.Async.cs	Uses a pooled buffer to collate dynamic header data without extra array allocations.
src/libraries/System.IO.Compression/src/System/IO/Compression/ZipArchive.cs	Central directory read buffer is rented from ArrayPool<byte> instead of allocating a new array per read.
src/libraries/System.IO.Compression/src/System/IO/Compression/ZipArchive.Async.cs	Async central directory parsing avoids range-operator array allocations by using ReadOnlyMemory<byte> slices over pooled buffers.
src/libraries/System.IO.Compression/src/System/IO/Compression/DeflateManaged/InflaterManaged.cs	Replaces small write loops with span-based fill/clear operations.