Add timeout option to all `run_*` functions

[rwaffen]

When calling a function like run_task from within a bolt plan, it cannot currently be interrupted from the plan level, nor is it possible to define a per-call timeout. The only available option is a global timeout.

However, it would be very useful to have finer control over this behavior. Ideally, timeouts could be defined per task, command, or sub-plan instead of relying solely on a single global timeout.

This would allow much more stable and reliable execution flows. With configurable timeouts, run_* functions could be aborted earlier when necessary, rather than having to wait until the global timeout is reached.

[marcusdots]

Can you make use of /usr/bin/timeout to achieve desired result?

$ /bin/bash -c "timeout 1 tail -f /dev/null && echo success || echo timeout"
timeout

[jay7x]

I opened #200 as another option to solve this.
But we should have _timeout option in run_* functions anyway.
Same as we have with catch_errors() function and _catch_errors option in run_* functions.

[jay7x]

btw, there is another trick:

$future = background() || { run_task('deploy', $target) }
# You can do something else meanwhile..
$future.wait($timeout) # This should raise Bolt::FutureTimeout error