Follow-ups for the background worker stack

Things left to do or decide after #2617 and its stack, so they don't get lost.

• Validate a configuration before stopping the running one: Start() calls Init() after the old app stopped, so any declaration error leaves the site down while Caddy rolls the configuration back. A side-effect-free preflight shared with Init(), called from Validate(). https://github.com/php/frankenphp/pull/2617#discussion_r3987787721

• Synchronize the startup-failure boundary: publishing ShuttingDown wakes the startup waiter before the thread finished its exit path, and a shutdown request is not distinguished from a thread that exited. Both worker handlers share that path. https://github.com/php/frankenphp/pull/2617#discussion_r3987787784

• Pull the rest of the common bookkeeping into workerLifecycle once that boundary has one shape. https://github.com/php/frankenphp/pull/2617#discussion_r3987787771

• num_regular_threads / max_regular_threads, so nobody computes the thread budget by hand. https://github.com/php/frankenphp/pull/2617#discussion_r3866053582

• Implement Io\Poll\Handle on FrankenPHP\WorkerHandle and the task handles, done in #2617 and #2636 on top of the 8.6 job of #2658

• Drop the per-task php_stream from the common path of #2636, now that a handle's lifetime is exactly the task's: hold the task state in the object, settle in free_obj, and build the stream lazily in getStream(), so a script that never multiplexes allocates no stream and no resource per task. Then recycle the handles themselves, through a per-thread free list. Channels stay eventfds on Linux; on macOS look for the equivalent, a kqueue descriptor triggered with EVFILT_USER is pollable and needs no pair, a pipe is the usual fallback (libuv's async handles), and Windows keeps the socket pair for php_select().

Start with #2617 and #2636, then read the linked discussions and the worker entry points named here: Start(), Init(), Validate(), workerLifecycle, and the worker/task handle paths. Choose one unchecked follow-up; it is done when that specific configuration, lifecycle, thread-budget, handle, or platform-channel behavior is implemented and the corresponding checklist item can be verified.