Decode URL-encoded zip entry names when extracting NuGet packages

[Copilot]

NuGet packages percent-encode file paths in their zip entries (e.g. spaces → %20). ZipFileExtensions.ExtractToDirectory does not decode these, resulting in files being written with names like file%20with%20spaces.txt on disk instead of file with spaces.txt.

Changes

• ModuleFast.psm1: Replace the single ExtractToDirectory call with a manual loop that applies [Uri]::UnescapeDataString to each entry's FullName before constructing the destination path, then extracts via ZipFileExtensions.ExtractToFile. Directory entries (both / and \ terminators) are handled explicitly.

foreach ($entry in $zip.Entries) {
    $decodedEntryName = [Uri]::UnescapeDataString($entry.FullName)
    $destPath = Join-Path $installPath $decodedEntryName
    # ZIP spec uses '/' but some tools emit '\'; treat both as directory entries
    if ($decodedEntryName.EndsWith('/') -or $decodedEntryName.EndsWith('\')) {
        New-Item -ItemType Directory -Path $destPath -Force | Out-Null
    } else {
        $destDir = [Path]::GetDirectoryName($destPath)
        if (-not (Test-Path $destDir)) { New-Item -ItemType Directory -Path $destDir -Force | Out-Null }
        [IO.Compression.ZipFileExtensions]::ExtractToFile($entry, $destPath, $true)
    }
}

• ModuleFast.tests.ps1: Unit test added that builds an in-memory ZIP with percent-encoded directory and file entries, runs the extraction logic, and asserts decoded paths exist while encoded paths do not.