On Tue, Jun 21, 2011 at 11:45 PM, Dave Reisner wrote:

On Wed, Jun 22, 2011 at 01:27:54AM -0300, Gerardo Exequiel Pozzi wrote:

...

By default, there is no network is configured in rc.conf. So if you boot the machine with an unmodified rc.conf you will see a warning about network deprecation settings.

Signed-off-by: Gerardo Exequiel Pozzi ---  rc.conf |    2 +-  1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/rc.conf b/rc.conf index 5984171..bd13fce 100644 --- a/rc.conf +++ b/rc.conf @@ -114,4 +114,4 @@ NETWORK_PERSIST="no"  # If something other takes care of your hardware clock (ntpd, dual-boot...)  # you should disable 'hwclock' here.  # -DAEMONS=(hwclock syslog-ng network netfs crond) +DAEMONS=(hwclock syslog-ng !network !netfs crond) -- 1.7.5.2

I noticed this the other day on one of my VMs. I'm not sure what the best course of action is. There's a few choices...

1) This. 2) Refactor the need_legacy check. It's fairly naive at the moment. We could easily cover the 'null' edge case by checking:

 [[ -z $INTERFACES && -z $interface ]]

3) Change where deprecation messages are printed -- i.e., only once we're about to execute some net-tools based logic do we print a message. This is perhaps the least desireable, imo, because it's also prone to creating the most amount of spammage. I wouldn't want to see the warning printed more than once per invocation.

I'd be wary of disabling network by default. I agree with Dave's #2 suggestion though. -Dan

On Wed, Jun 22, 2011 at 6:47 PM, Thomas Bächler wrote:

Am 22.06.2011 18:17, schrieb Dan McGee:

...

On Tue, Jun 21, 2011 at 11:45 PM, Dave Reisner wrote:

...

On Wed, Jun 22, 2011 at 01:27:54AM -0300, Gerardo Exequiel Pozzi wrote:

...

By default, there is no network is configured in rc.conf. So if you boot the machine with an unmodified rc.conf you will see a warning about network deprecation settings.

Signed-off-by: Gerardo Exequiel Pozzi ---  rc.conf |    2 +-  1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/rc.conf b/rc.conf index 5984171..bd13fce 100644 --- a/rc.conf +++ b/rc.conf @@ -114,4 +114,4 @@ NETWORK_PERSIST="no"  # If something other takes care of your hardware clock (ntpd, dual-boot...)  # you should disable 'hwclock' here.  # -DAEMONS=(hwclock syslog-ng network netfs crond) +DAEMONS=(hwclock syslog-ng !network !netfs crond) -- 1.7.5.2

I noticed this the other day on one of my VMs. I'm not sure what the best course of action is. There's a few choices...

1) This. 2) Refactor the need_legacy check. It's fairly naive at the moment. We could easily cover the 'null' edge case by checking:

 [[ -z $INTERFACES && -z $interface ]]

3) Change where deprecation messages are printed -- i.e., only once we're about to execute some net-tools based logic do we print a message. This is perhaps the least desireable, imo, because it's also prone to creating the most amount of spammage. I wouldn't want to see the warning printed more than once per invocation.

I'd be wary of disabling network by default. I agree with Dave's #2 suggestion though.

ACK.

Yup, number two makes most sense to me too. -t