Correctly set FileOffsets in WriteHandle

[OneBlue]

Summary of the Pull Request

This change updates WriteHandle to correctly set offsets when writing to its handle.

If the handle was a file, then WriteHandle would always write to the same offset, effectively incorrectly truncating the file

PR Checklist

• Closes: Link to issue #xxx
• Communication: I've discussed this with core contributors already. If work hasn't been agreed, this work might be rejected
• Tests: Added/updated if needed and all pass
• Localization: All end user facing strings can be localized
• Dev docs: Added/updated if needed
• Documentation updated: If checked, please file a pull request on our docs repo and link it here: #xxx

Detailed Description of the Pull Request / Additional comments

Validation Steps Performed

[Copilot]

Pull request overview

This PR fixes wsl::windows::common::relay::WriteHandle so that when the underlying handle is a seekable file, overlapped writes advance the file offset instead of repeatedly writing at the same position (which could overwrite/truncate previously written data).

Changes:

• Add per-handle offset tracking to WriteHandle and apply it to OVERLAPPED.Offset/OffsetHigh for each write.
• Advance the tracked offset on both synchronous and asynchronous write completion paths.
• Add a new Windows test that validates WriteHandle behavior for pipes and disk files.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
src/windows/common/relay.hpp	Adds stored offset state to WriteHandle.
src/windows/common/relay.cpp	Initializes and updates WriteHandle’s file offset and sets OVERLAPPED offsets before each write.
test/windows/WSLCTests.cpp	Adds WriteHandleContent test covering pipe and file output scenarios.

[Copilot]

The ReadHandle callback appends buffer.data() even when buffer can be an empty span (e.g., EOF/broken pipe paths may invoke the callback with an empty span whose .data() is null). Passing a null pointer to std::string::append(ptr, 0) is undefined behavior and can crash under some CRT/debug configurations. Guard on buffer.empty() (or append only when buffer.size() > 0).

Suggested change

	std::move(readPipe), [&](const gsl::span<char>& buffer) { readData.append(buffer.data(), buffer.size()); }));
	std::move(readPipe),
	[&](const gsl::span<char>& buffer)
	{
	if (!buffer.empty())
	{
	readData.append(buffer.data(), buffer.size());
	}
	}));

[Copilot]

This test creates a fixed relative filename ("write-handle-test") in the current working directory and uses CREATE_NEW semantics. If a previous run left the file behind (crash/abort) or tests run concurrently in the same working dir, this can become flaky. Prefer creating the file under std::filesystem::temp_directory_path() with a unique name (and/or do a best-effort pre-cleanup) and remove via the same full path.

Suggested change

	auto outputFile = wil::create_new_file(L"write-handle-test", GENERIC_WRITE | GENERIC_READ, 0, nullptr);
	auto cleanup = wil::scope_exit_log(WI_DIAGNOSTICS_INFO, [&]() {
	outputFile.reset();
	std::filesystem::remove("write-handle-test");
	std::filesystem::path outputFilePath =
	std::filesystem::temp_directory_path() /
	(L"write-handle-test-" + std::to_wstring(GetCurrentProcessId()) + L".tmp");
	auto outputFile = wil::create_new_file(outputFilePath.c_str(), GENERIC_WRITE | GENERIC_READ, 0, nullptr);
	auto cleanup = wil::scope_exit_log(WI_DIAGNOSTICS_INFO, [&]() {
	outputFile.reset();
	std::error_code ec;
	std::filesystem::remove(outputFilePath, ec);