Syntax/invocation of rc.local vs. rc.local_shutdown

moosejaw · 01-29-2016, 02:15 PM

I'm curious why, in the Slackware rc.d scripts, the local startup and shutdown scripts (/etc/rc.d/rc.local and /etc/rc.d/rc.local_shutdown, respectively) are invoked differently:

rc.local is invoked without an argument: when it is called in /etc/rc.d/rc.M, the syntax is
Code:
```
. /etc/rc.d/rc.local
```
rc.local_shutdown is invoked with a "stop" argument: in /etc/rc.d/rc.{6,K}, the syntax is
Code:
```
/etc/rc.d/rc.local_shutdown stop
```

This entails some differences in how one writes rc.local and rc.local_shutdown (though rc.local_shutdown will happily ignore the "stop" parameter if you don't tell it to do anything with it, as far as I can tell).

This isn't a problem, obviously, but I'm just curious why things are set up in this way. Why the disparity in how these two scripts are handled by the system?

willysr · 01-29-2016, 02:27 PM

From BASH Howto:

Quote:

When a file is sourced (by typing either source filename or . filename at the command line), the lines of code in the file are executed as if they were printed at the command line. This is particularly useful with complex prompts, to allow them to be stored in files and called up by sourcing the file they are in.

GazL · 01-29-2016, 02:52 PM

There's a lot of inconsistency in rc.S and rc.M relating to launching the other rc.* files.
Some are sourced, some called with "sh rc.something", and some are executed directly. Personally, I'd prefer them all to be invoked directly as /etc/rc.d/rc.something (if only for the sake of consistency) and left to rely on the "#!" line to decide what shell they run in. I actually went through it all on 14.1 at one point and converted them to run that way. Didn't see any adverse effects, but that was a while ago and I don't have the changes any more.

In the old days, one would source rather than exec to avoid the fork/exec and process setup/teardown overhead associated with spawning a new shell process, but with today's hardware the speed saving is unlikely to be noticeable and its my belief that the additional reliability of running independent shells isolated from each other rather than sourcing them would outweigh it.

The inconsistency is a bit ugly, but as you say, it's not really hurting anything (other than being something systemd proponents can point at and go "EWWWWWWWWWWWW!"

). I'd be happy to see it changed, but if its not, I wouldn't make a fuss.

moosejaw · 01-30-2016, 08:42 PM

Thanks both. I get that sourcing is useful for complex prompts, as the bash howto says. It seems to me it would be equally useful for /etc/rc.d/rc.local_shutdown, which is just as likely as /etc/rc.d/rc.local to consist of complex prompts. Anyway, again, not a big deal, but thanks for indulging my curiosity here.