On 05/18/2011 01:27 PM, Vratislav Podzimek wrote:
These are two patches that povides WPA wireless functionality both as command line options and in kickstart. They also change the way we communicate with the Network Manager when creating wireless connections (all WPA, WEP and unprotected) to use D-Bus.
The integration into loader network configuration paths looks reasonable to me (given their poor state). Some things apart from what Chris and Peter remarked: 1) Is it possible to let NM generate UUID for the new connection by not setting the value? We don't generate connection UUID anywhere in acaconda and I'd like to keep this commitment. + + connection = nm_connection_new(); + + s_con = (NMSettingConnection*) nm_setting_connection_new(); + uuid = nm_utils_uuid_generate(); + g_object_set (G_OBJECT (s_con), + NM_SETTING_CONNECTION_UUID, uuid, ... just by leaving out this setting + NM_SETTING_CONNECTION_ID, ssid, + NM_SETTING_CONNECTION_TYPE, "802-11-wireless", + NULL); + g_free (uuid); + nm_connection_add_setting (connection, NM_SETTING (s_con)); 2) We should remove our handling of keyfiles from loader (creating in writeEnabledNetInfo), and unify behaviour for wep and wpa in anaconda (stage2), e.g. copy also wpa keyfiles to target system - instead of writeWepkeyFile, we would probably just copy keyfiles we find. Also in kickstart stage2 processing, we should either stop writing out wep key file (this would happen in the case when the device is not activated in loader), or write out also wpa keys. I'm for the former. 3) Should we mention something about the behaviour of options in documentation? Or perhaps document it at least internally. For example behaviour regarding device specified/unspecified vs. ssid option present. Radek _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list
