Lokad.Awk 0.4.0

Lokad.Awk

An embeddable awk interpreter for .NET 10. Run ordinary text-processing programs while your application controls files, streams, external commands and resource admission.

dotnet add package Lokad.Awk

Embedding

Supply command-line-style arguments and an IAwkHost implementation. All IO passes through the host; the package does not provide a filesystem or process host.

using Lokad.Awk;

var invocation = AwkCommandInvocation.CreateWithStandardDescriptors(
    "awk",
    ["{ total += $2 } END { print total }", "/data/sales.tsv"],
    [new AwkEnvironmentVariable("PWD", "/")]);

if (!Awk.TryParse(invocation, out var awk))
    throw new InvalidOperationException("Not an awk command.");

IAwkHost host = new ApplicationAwkHost(); // Implement the application's IO boundary.
var policy = AwkExecutionPolicy.CreateRestricted()
    .WithAccessiblePathRoot(new AwkPath("/data"));
using var cancellation = new CancellationTokenSource(TimeSpan.FromSeconds(10));
var exitCode = await awk.ExecuteAsync(host, policy, cancellation.Token);

TryParse recognizes the invocation; execution loads, parses and runs the program. ApplicationAwkHost above is application code; the test suite's MockFileSystem illustrates a host implementation.

Use caller cancellation for deadlines and WithMemoryBudget(IMemoryBudget) for shared memory admission. The restricted preset limits capabilities, stack depth and open resources; aggregate memory admission requires a supplied budget. Text must be valid UTF-8. See the runtime contract for host ownership, cleanup and protection guarantees.

Regex and compatibility

Regex operations require a host-supplied IAwkRegexEngine, configured with policy.WithRegexEngine(engine). Core has no regex backend or Utf8Regex dependency. A copyable adapter uses published Lokad.Utf8Regex 0.4.0 with cooperative cancellation. Its POSIX longest-match and capture-selection gaps remain documented limitations. Non-regex processing needs no provider.

Supported core behavior includes records/fields, expressions, loops, functions, flat arrays, formatting and getline, with selected GNU conveniences. This is not full POSIX/GNU conformance: CSV mode, nested arrays, system() and namespaces remain unsupported.

• Compatibility: supported boundaries and known differences.
• Benchmarks: representative elapsed-time ratios and managed allocations.
• 0.4.0 migration notes: breaking host, regex and policy API changes.

Before 1.0, minor releases may remove public APIs as described in the release notes.

Building

Use the stable SDK pinned in global.json, then run dotnet test Lokad.Awk.slnx -c Release. CI builds, tests and packs on Windows and Linux. Only src/Lokad.Awk produces a NuGet package; the adapter example, tests and benchmarks are not packaged.