Re: New behavior proposal --halt -% with job killing

[Ole Tange]

On Mon, Apr 27, 2015 at 5:46 PM, Martin d'Anjou
<martin.danjou14@gmail.com> wrote:
:
> halt_argument_value: "never" | (("onsuccess" | "onfail") "," quantity
> "," consequence)
> quantity: number | pct
> number: [1-9][0-9]*
> pct:[1-9][0-9] "%"
> consequence: "killpending" | "killall"
:
> onsuccess,N,consequence
>     Halt after N jobs run sucessfully.
>
>     The number N ranges from 1 to the total number of jobs, or a
> percentage. If N is a percentage (N%), then at least 3 jobs have run
> to completion before GNU Parallel considers making a decision. GNU
> parallel will halt when the percentage of successful jobs will equal
> or exceed N%.
>
>     If the consequence is "killpending", GNU parallel will wait for
> all running jobs to come to completion, but no more new jobs will be
> dispatched (the pending jobs will be killed).
>
>     If the consequence is "killall", GNU parallel will stop
> dispatching jobs and will kill all the running jobs.

I think we are pretty much in agreement. I just prefer calling
'killall' for '--halt now' and 'killpending' for '--halt soon'.

So:

halt soon if failed jobs = 20%: --halt soon,fail,20% (if 20% have
failed, start no new jobs and wait until the running jobs have
finished).

halt now if number of successes = 3: --halt now,success,3 (if 3 jobs
succeeded, start no new jobs and kill off the running jobs now).

halt never: --halt 0 == --halt never

The reason I want N > 2 is to avoid confusion with existing (magic)
numbers, and so that order does not matter:

soon,fail,20 == 20,fail,soon

But I can easily imagine that N=1 would be nice to have: Of the 10000
solutions find one that works and stop. Maybe by simply requiring
giving all 3 parameters?

And I will suggest that you can use mix , and = as separators, so you
can write it as:

--halt now,fail=1 == --halt now,fail,1

> Backward compatibility:
>
> --halt 0 is the same as --halt never
>
> --halt 1 is the same as --halt onfail,1,killpending

--halt soon,fail=1

> --halt 2 has no equivalents (it does not make sense to orphan jobs)

But --halt 2 should probably killall anyway.

--halt now,fail=1

> --halt -1 is the same as --halt onsuccess,1,killpending

--halt soon,success=1

> --halt -2 has no equivalents (it does not make sense to orphan jobs)

But --halt -2 should probably killall anyway.

--halt now,success=1

> --halt 1-99% is equivalent to --halt onfail,1-99%,killpending

--halt soon,fail=1-99%


I think we are close to a decent design, and I will see how hard that
is to implement.

/Ole