iron@... (Mike Orr) wrote:
>Sorry, I wasn't following the thread that closely.
>init.d/functions is definitely not a standard. The init.d/ directory itself
>exists only on Unices that use the System V style initialization system.
>In the BSD style, each service is started by a stanza in a general startup
>script (/etc/rc.d/rc.M in Slackware Linux), not by its own script.
>Nevertheless, Slackware supports an init.d/ directory for compatibility
>with third-party packages.
I don't remember the different standards, but for what it's worth, Solaris has
/etc/init.d and /etc/rc*.d for the different scripts. The former is where
actual scripts tend to be placed, and the latter directories usually contain
symbolic links to the former. On Red Hat 6.1, there's /etc/rc.d/init.d and
/etc/rc.d/rc*.d, with the same kind of symbolic linking conventions, although
it's wiser to use chkconfig on RH6.1 to manage these, of course.
>My Debian Linux uses init.d/, but there's no init.d/functions .
It's a Red Hat thing.
>Also, the location of the init.d/ directory is all over the map.
>Red Hat Linux uses /etc/rc.d/init.d/, Debian Linux uses /etc/init.d/,
>and Jeff's system uses /usr/etc/rc.d/init.d/ . The location of the
>startup scripts in one of the biggest inconsistencies between Unices.
I think Debian follows SysV most closely, then. Solaris' documentation mentions
the motivations for changing the directory hierarchy - I think they arranged it
slightly differently in SunOS 4.x and earlier.
>Which functions do we need? Attached is a copy of the functions file
>from Red Hat 6.2. Of course, some of the code is specific to that
This is probably the script being discussed. It might be wise to assess the
portability of it - lots of Linux-related stuff tends to assume the presence of
various GNU programs which might not be there on other UNIX systems.
Get your firstname@... email for FREE at http://Nameplanet.com/?su