[PATCH] run-tests: "fix" race condition in race condition fix

[Pierre-Yves David]

On 01/22/2016 12:08 PM, timeless wrote:
> Pierre-Yves David wrote:
>> How much of your issue is related to "bryan patch failure to fix the race
>> introduced by Laurent" and how much is it of "run-tests.py interruption have
>> been borked since forever"?
>
> all of it.
>
>> I've personnally never have been confortable with ctrl^C run-tests.py in years.
>
> I've always been moderately annoyed by it, but I do get a prompt if I
> don't try to fix the errors. If I or anyone else (i.e. bos) tries to
> fix the channel stuff, the result tends to be no prompt because
> something swallows something and the script doesn't terminate, which
> is definitely worse than some ugly exceptions.

What's the way forward here, is the Bryan approach adjustable to provide 
good result? Do you have an alternative in mind? should we backout 
Laurent's patch?

-- 
Pierre-Yves David