[arch-releng] arch-releng is working

Aaron Griffin aaronmgriffin at gmail.com
Tue Jan 20 14:10:55 EST 2009


Moving this to the releng list

On Tue, Jan 20, 2009 at 12:57 PM, Gerhard Brauer <gerbra at archlinux.de> wrote:
> Am Tue, 20 Jan 2009 10:02:59 -0600
> schrieb Aaron Griffin <aaronmgriffin at gmail.com>:
>> I don't know what the fix is. I don't think I saw a clear definition
>> of the problem. What's the issue?
>
> I've made a error in installer changes.
> You should revert this commit:
> http://projects.archlinux.org/?p=installer.git;a=commit;h=d8510783cac2fd8b1837e3481ef04ddde8adaafb
>
> Don't know exactly what i have drunken, as i checked the patch it
> worked.... And it doesn't work without the patch....
>
> Here is a script which demonstrates that the original code in installer
> was ok:
> ---------
> #!/bin/sh
> #
>
> cp /tmp/x/rc.conf.org /tmp/x/rc.conf
>
> BROADCAST="10.0.0.255"
> IPADDR="10.0.0.20"
> SUBNET="255.255.255.0"
> INTERFACE="eth1"
>
> sed -i "s#eth0=\"eth0#$INTERFACE=\"$INTERFACE#g" /tmp/x/rc.conf
> sed -i "s# 192.168.0.2 # $IPADDR #g" /tmp/x/rc.conf
> sed -i "s# 255.255.255.0 # $SUBNET #g" /tmp/x/rc.conf
> sed -i "s# 192.168.0.255\"# $BROADCAST\"#g" /tmp/x/rc.conf
> sed -i "s#eth0)#$INTERFACE)#g" /tmp/x/rc.conf
> cp /tmp/x/rc.conf /tmp/x/rc.conf.good
>
>
> cp /tmp/x/rc.conf.org /tmp/x/rc.conf
> sed -i "s#eth0=\"eth0#$INTERFACE=\"$INTERFACE#g" /tmp/x/rc.conf
> sed -i "s#192.168.0.2# $IPADDR#g" /tmp/x/rc.conf
> sed -i "s#255.255.255.0# $SUBNET#g" /tmp/x/rc.conf
> sed -i "s#192.168.0.255\"#$BROADCAST\"#g" /tmp/x/rc.conf
> sed -i "s#eth0)#$INTERFACE)#g" /tmp/x/rc.conf
> cp /tmp/x/rc.conf /tmp/x/rc.conf.bad
> --------
>
> Output for network settings from rc.conf.good (the original code):
> eth1="eth1 10.0.0.20 netmask 255.255.255.0 broadcast 10.0.0.255"
> INTERFACES=(eth1)
>
> Output for network settings from rc.conf.bad (my patch):
> eth1="eth1  10.0.0.20 netmask  255.255.255.0 broadcast  10.0.0.2055"
> INTERFACES=(eth1)
>
> The additional spaces are not the point, but the patched code generates
> a bad broadcast.

OK so, the original code works fine and that commit should be reverted
completely? Or does it still need fixing?

> BTW: I've posted already on arch-releng. Do you think we could do the
> release/finish on/around friday?

Yes. I just asked the developers to have core frozen and stable on
Friday so we can build the release.

Notes:
* My build-and-upload script failed last night. Whoops. Rebuilding now.
* Thomas recommended we remove madwifi completely. Done
* We need to write a small README explaining each image and
recommending people do ftp installs unless they need the core packages
for offline installation


More information about the arch-releng mailing list