Add net11 Process.ReadAllLines

[SimonCropp]

Adds the synchronous counterpart to the already polyfilled ReadAllLinesAsync:

IEnumerable<ProcessOutputLine> ReadAllLines(TimeSpan? timeout = null)

API count 1072 to 1073. This was the last item outstanding from the net11 RC1 sweep.

Semantics

Taken from net11 by driving a child process that writes to both streams with delays between them, rather than assumed:

• the sequence is lazy. The call returned in 0ms and lines were yielded at 51ms, 755ms and 1458ms as the child produced them, rather than being collected first
• standard output and standard error are interleaved in arrival order
• exceeding the timeout throws TimeoutException, matching WaitForExitOrThrow and the other synchronous helpers on Process
• a process whose output is not redirected throws InvalidOperationException

Implementation

An iterator over the same ConcurrentQueue and SemaphoreSlim arrangement ReadAllLinesAsync already uses, with a blocking wait bounded by the remaining timeout.

Because an iterator body does not run until it is enumerated, the timeout is measured from the point enumeration starts rather than from the call. Recorded as a //Note:.

One gap in the tests

Nothing asserts the TimeoutException path. Provoking it needs a child that stays quiet for a known interval, and every candidate available to the test project either exits promptly or makes the assertion depend on timing, which would be flaky on CI. The path was exercised by hand against net11 with a purpose built child instead. For the same reason there is no test asserting laziness, though that was verified the same way.

The timeout parameter is still covered, with a value large enough not to fire, so the remaining-time arithmetic is exercised.

Verification

• Solution builds clean in Release. Consume builds clean in Debug across all 22 target frameworks.
• Tests green on net11.0 (1672), net462 (1638), net8.0 (1669) and net10.0 (1672), plus PublicTests, EmbeddedTests, UnsafeTests, NoRefsTests and NoExtrasTests.
• Since these tests start real processes, the three new ones were run repeatedly on both net11.0 and net462 to check for timing flakiness. No failures.