patch, 2 partsgit-p4: kill watchdog and suppress irrelevant output

> From: Lars Schneider <larsxschneider@gmail.com>
>
> If tests are executed in verbose mode then the retry logic clutters the
> test output. Suppress that clutter.
>
> Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
> ---
>  t/lib-git-p4.sh | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/t/lib-git-p4.sh b/t/lib-git-p4.sh
> index 30bf7ae..03f29c1 100644
> --- a/t/lib-git-p4.sh
> +++ b/t/lib-git-p4.sh
> @@ -174,7 +174,7 @@ retry_until_fail() {
>  	until ! "$@" 2>/dev/null || test $(time_in_seconds) -gt $timeout
>  	do
>  		sleep 1
> -	done
> +	done >/dev/null 2>&1

> larsxschneider@gmail.com writes:
> 
>> From: Lars Schneider <larsxschneider@gmail.com>
>> 
>> If tests are executed in verbose mode then the retry logic clutters the
>> test output. Suppress that clutter.
>> 
>> Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
>> ---
>> t/lib-git-p4.sh | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>> 
>> diff --git a/t/lib-git-p4.sh b/t/lib-git-p4.sh
>> index 30bf7ae..03f29c1 100644
>> --- a/t/lib-git-p4.sh
>> +++ b/t/lib-git-p4.sh
>> @@ -174,7 +174,7 @@ retry_until_fail() {
>> 	until ! "$@" 2>/dev/null || test $(time_in_seconds) -gt $timeout
>> 	do
>> 		sleep 1
>> -	done
>> +	done >/dev/null 2>&1
> 
> Eh, what does this squelch?  The sleep in the body of the loop is
> silent, "test A -gt B" on the loop condition would be silent too, so
> you are squelching the invocation of "$@" whose standard error
> stream is already sent to 2>/dev/null?
> 
> If so, why not do it there instead?  You seem to run only "kill" to
> send some signal to a process using this helper function, and it
> would be silent on its standard output stream (even though it may
> say "no such process" etc. on its standard error), so it is not
> clear to me what you are doing with this change here...

>>
>> If so, why not do it there instead?  You seem to run only "kill" to
>> send some signal to a process using this helper function, and it
>> would be silent on its standard output stream (even though it may
>> say "no such process" etc. on its standard error), so it is not
>> clear to me what you are doing with this change here...
>
> If I run git-p4 tests in verbose mode (e.g. "./t9823-git-p4-mock-lfs.sh -v") without this patch then the last lines of the output look like this:
>
>>>> Output Start >>>
> expecting success:
> 	kill_p4d
>
> ./lib-git-p4.sh: line 172: 26289 Killed: 9               while true; do
>      if test $(time_in_seconds) -gt $timeout; then
>          kill -9 $pid; exit 1;
>      fi; sleep 1;
> done
> ok 8 - kill p4d
>
> # passed all 8 test(s)
> 1..8
> <<< Output end <<<
>
> However, I want them to look like this:
>
>>>> Output Start >>>
> expecting success:
> 	kill_p4d
>
> ok 8 - kill p4d
>
> # passed all 8 test(s)
> 1..8
> <<< Output end <<<
>
> This is achieved with the patch. I am no shell expert ... is there a nicer way to achieve the same?

> From: Lars Schneider <larsxschneider@gmail.com>
>
> Hi,
>
> these patches extend "git-p4: add trap to kill p4d on test exit" (dfe90e8)
> and therefore should be applied on master.

>
> Thanks,
> Lars
>
> Lars Schneider (2):
>   git-p4: kill p4d watchdog on cleanup
>   git-p4: suppress non test relevant output
>
>  t/lib-git-p4.sh | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> --
> 2.5.1

> larsxschneider@gmail.com writes:
> 
>> From: Lars Schneider <larsxschneider@gmail.com>
>> 
>> Hi,
>> 
>> these patches extend "git-p4: add trap to kill p4d on test exit" (dfe90e8)
>> and therefore should be applied on master.
> 
> Wait, wait.  Please be a bit more careful when you use such a
> phrasing.  Did somebody review these and said that these are
> trivially correct improvements?
> 
> This depends on what you exactly meant "extend" and "improve" (for
> the other one); if the "improvement" were to make something that
> used to be unusable to usable, then the more sensible way forward
> during the -rc stabilization period might be to revert the earlier
> merges to 'master' that brought in undercooked topic.