Fix bulk scheduling in stdexec::parallel_scheduler

[Cra3z]

Currently, the following code fails to compile (See https://godbolt.org/z/W5ah3Y3cG):

auto sched = stdexec::get_parallel_scheduler();
auto d = stdexec::get_completion_domain<>(stdexec::get_env(stdexec::schedule(sched)), stdexec::env<>{});
static_assert(std::same_as<decltype(d), stdexec::__parallel_scheduler_domain>);

The completion domain of parallel_scheduler's sender should be stdexec::__parallel_scheduler_domain. parallel_scheduler should be able to query get_completion_domain_t<set_value_t> instead of get_domain_t.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.