154862 2006-11-12 02:12 0000 baselayout-1.13.0_alpha5,6,7 changed the location wpa_supplicant.conf is read from without mention 2006-11-12 06:14:26 0000 1 1 1 Unclassified Gentoo Linux baselayout unspecified All Linux RESOLVED FIXED P2 trivial --- 1 ikelos@gentoo.org base-system@gentoo.org ikelos@gentoo.org 2006-11-12 02:12:55 0000 Baselayout-1.13.0_alpha5,6,7 now read wpa_supplicant.conf from /etc/wpa_supplicant/wpa_supplicant.conf, rather than the /etc/wpa_supplicant.conf which previously used to work. People using WPA will then lose their connection without much indication of the problem on reboot. Whilst /etc/conf.d/net.example does offer the answer, it wasn't obvious that it was a wpa_supplicant issue to being with. The simplest solution seems like a small einfo/ewarn during postinst... uberlord@gentoo.org 2006-11-12 02:37:31 0000 Why not edit /lib/rcscripts/net/wpa_supplicant.conf and read lines 220-226. You'll see we use config file locations as follows /etc/wpa_supplicant/wpa_supplicant-$iface.conf /etc/wpa_supplicant/wpa_supplicant.conf /etc/wpa_supplicant.conf We use the first one we find. wpa_supplicant has never installed /etc/wpa_supplicant/wpa_supplicant.conf, so if it's being used then you created it. This is hardly a bug. ikelos@gentoo.org 2006-11-12 03:05:39 0000 Sorry about that Roy, I've been using wpa_supplicant-0.5.5 since September when it was released, but hadn't realised that it installs an /etc/wpa_supplicant/wpa_supplicant.conf file by default, seems you missed it too. With the previous baselayout this wasn't ever a problem, but since it now causes a problem, I'm not sure how best to fix it other than a postinst warning. Can you think of any other solutions? uberlord@gentoo.org 2006-11-12 06:14:26 0000 (In reply to comment #2) > Sorry about that Roy, I've been using wpa_supplicant-0.5.5 since September when > it was released, but hadn't realised that it installs an > /etc/wpa_supplicant/wpa_supplicant.conf file by default, seems you missed it > too. With the previous baselayout this wasn't ever a problem, but since it now > causes a problem, I'm not sure how best to fix it other than a postinst > warning. You're right, I forgot about that. I'll put a notice in the wpa_supplicant ebuild Fixed in wpa_supplicant-0.5.5