Initscripts and tput

[Nathaniel McCallum]

On Wed, 2004-09-01 at 15:12 -0700, Matt Zimmerman wrote:
> On Wed, Sep 01, 2004 at 01:12:08PM +0100, Scott James Remnant wrote:
> 
> > I think I prefer this alternative though:
> > 
> >    [ -x /usr/bin/tput ] && tput ...
> > 
> > Colour becomes "active" once /usr has been mounted, which is relatively
> > early in the boot sequence anyway -- most of the preceding scripts won't
> > need OK/FAIL type messages.  And for 95% of installs, this will work
> > from script#1 because /usr won't be separately mounted.
> 
> Right, the only stuff which happens before mountall.sh is:
> 
> S01devfsd
> S02mountvirtfs
> S04udev
> S05bootlogd
> S05initrd-tools.sh
> S05keymap.sh
> S07hdparm
> S09scsitools-pre.sh
> S10checkroot.sh
> S15isapnp
> S18hwclockfirst.sh
> S18ifupdown-clean
> S20module-init-tools
> S20modutils
> S22scsitools.sh
> S25libdevmapper1.00
> S25raid2
> S27evms
> S30checkfs.sh
> S30etc-setserial
> S30mt-st
> S30procps.sh
> 
> It's a long list, but there really isn't much to talk about during that
> time, and it's the stuff that we want to keep as simple as possible.  I
> don't think I mind if it doesn't look fancy, though we might want to quiet
> it down a bit.  It would be easy if bootlogd worked reliably. :-/
> 
> Thoughts?

Though, it would be nice to keep the function calls in those scripts
because we will use those function calls in usplash to provide status
updates.  Especially that early on (status updates for fsck etc).

One thing to consider: the shell function library that we are using is a
slight derrivitive of the lsb init-functions library from debian.  The
lsb one from debian uses both color and positional asci codes *and* they
are hardcoded.  For us to fall back to hardcoded if tput is not
available seems still an improvement from the debian script.

Nathaniel