Bootscript patches

Dan Nicholson dbn.lists at
Sat Jun 16 12:58:11 PDT 2007

On 6/16/07, Matthew Burgess <matthew at> wrote:
> On Saturday 16 June 2007 19:32:17 Dan Nicholson wrote:
> > Don't suppress stderr for some utilities. This can hide critical errors
> > during boot. For instance, I had an issue where mountfs was returning
> > with critical errors, but I couldn't see this issue. It turns out it was
> > just a missing mountpoint in /etc/fstab. This would have been easy to
> > spot of stderr from mount hadn't been suppressed.
> There's another couple of instances in console and the if{down,up} scripts.
> There may be reasons why these were left alone - I've only done a simple
> grep, not analysed them yet.  Also, I'd prefer the instances in modules were
> removed in this patch, rather than the cleanup patch - that way all the
> stderr suppression removal is in the same place.  That's really picking nits
> though.

I looked at console. In my testing, you get normal messages on stderr
with loadkeys, so I kept those in to keep the messaging consistent.
For ifup/ifdown, I think we want to suppress stderr for "ip link show
$1" in the "if" statement since it's just being used as a test for the
interface's existence. When the output is saved in link_status,
though, we might just want to let stderr go to the screen.

Yeah, I forgot about the 2>&1 in modules at first. I'll split them up
more appropriately.


More information about the lfs-dev mailing list