Re: Silent or move to /dev/null alsactl restore out

"Laurent Bercot" <[email protected]> writes:

>  I don't think it is the case, no. The myth started well before 
>  s6-rc
> was a thing. From the first days of s6, people assumed that, 
> despite
> the documentation explicitly saying the opposite; this has 
> nothing to
> do with execline's use in s6-rc.

Ah okay, fair enough. Perhaps we just need those writing 
guides/tutorials to emphasise that, when presenting examples of 
execline-based service scripts, that such scripts don't _need_ to 
be execline based. Because although these guides/tutorials might 
not explicitly say "you need to learn execline", the fact that 
people don't state up-front that execline is _not_ compulsory 
might lead people to assume that it _is_.

In the case of documentation i'm putting together, i'm making an 
explicit choice to have a shebang line with dash, because i don't 
want people to feel they have to grok execline before they can use 
s6 or s6-rc.

At any rate, i can only imagine your frustration at the longevity 
of this 'furphy', to use an Australianism:

  https://en.wiktionary.org/wiki/furphy#Noun

>  The point of hardcoding execline here is that I need a middle 
>  ground
> between "force up/down to be a path to an external script", 
> which is
> not flexible wrt e.g. existing sysvinit/openrc scripts you way 
> want to
> call with "start" and "stop" arguments, and "make up/down full
> scripts/executable files", which requires embedding 
> scripts/executables
> in the compiled db, which, no.
>
>  And that's *exactly* what execlineb is for.
>  The idea is for people to write
>     echo "/path/to/script start" > $oneshot/up
> without wondering too much about it, and *without* needing to 
> get lost
> in the details of execline syntax. But obviously, I suck at 
> predicting
> user behaviour, and people always end up wanting to be too smart 
> and
> tying their brains in knots about incredibly simple stuff, so, 
> whatever.

Heh. Well, i certainly wasn't trying to suggest that you'd made a 
_bad_ or _problematic_ design choice! Just wondering if the design 
choice had contributed to the "you must use execline" thing. But 
if, as you say, this thing has long preceded s6-rc, then indeed, 
the workings of s6-rc can't be a factor.


Alexis.