Hi,
On Mon, Aug 27, 2012 at 12:51:04AM +0200, intrigeri wrote:
> Hi,
>
> Ague Mille wrote (26 Aug 2012 09:30:19 GMT) :
> > While testing 0.13~rc1, I worked a bit on fixing the 'boot pause'
> > issue that is related to our readahead mechanism.
>
> Great!
>
> > The patch looks simple enough to be a candidate for 0.13.
>
> As you know, I tend to think the problem fixed by this patch is not
> worth the risks involved at this stage of the release process, but my
> extra-carefulness does not extend to vetoing if there is a general
> agreement with applying this patch before rc2.
I don't have strong opinion on this too. It might be a bit too late to
include it in the next release, but given how well the release tests are
going at the moment, we could maybe spend a bit more time on the next RC's
release tests to ensure this changes don't break something. Depends on the
number of people we'll be able to gather to do this tests I guess. How
heavily did you test this change already?
> > + start-stop-daemon \
> > + --start --background --pid /var/run/background-readahead.pid --startas /bin/sh -- \
> > + -c "$BG_FILES | xargs cat >/dev/null 2>&1")
>
> I assume you wanted to write --pidfile, as --pid does not exist in my
> copy of start-stop-daemon. Beware before s/pid/pidfile/, though:
> given --pidfile presence/absence changes quite drastically the
> behavior of start-stop-daemon, if the proposed patch works, then this
> option is probably not needed, is it?
>
> To end with, I have to say like the start-stop-daemon solution a lot.
Me too.
At the very least, this change should deserve a ticket on the wiki, as
defined in our process, even if it means closing it quickly.
bert.
