sequence of associating to a new WAP using PHP and DBus

don don
Mon Jul 5 07:35:20 PDT 2010

On Mon, 5 Jul 2010 10:01:21 +0200, Holger Schurig
<holgerschurig at> wrote:
> Hi !

Thank you for guiding  me, Holger!

>> However, an iwconfig gives me an
> While not related to your real-problem, you're doing something
> fundamentally 
> broken: basing a new design an WEXT and "iwconfig".
> iwconfig for example cannot really tell you if you're connected or not.
> plays tricks by giving back a wrong SSID or a wrong BSSID, but even that
> works 
> different when you use different cards.
> Use nl80211 instead, both in your kernel and in wpa_supplicant. And use
> "iw" 
> instead of "iwconfig" & friends.
Okay, thanks for the heads up!
>> occasionally. It seems obvious that my values are getting accepted but
>> not
>> validated by the addNetwork(), but they are causing selectNetwork() to
>> choke.
> Did you try this:
> a) run wpa_supplicant with a .conf file
> b) use wpa_cli to display the network blocks
> c) now use you DBUS code to make the changes you want
> d) again use wpa_cli to display the network blocks again
>    (maybe you use "get_network", but if you think the PSK hasn't
>    been transferred correctly, you might want to try "save_config"
>    and look at the newly written config file)
> If the network blocks aren't what you expected, change you code and
> at 
> c). You now know at least what piece of information is missing.

Yes... all of it. :) After figuring out the bit about Debian builds
killing wpa_cli, I was able to discover that I'm getting three new empty
networks that default to disabled. Thanks especially for that tip on
save_config, my man page does not mention it at all.

I'm still working on it, but I wanted to reply to thank you before you
exit for the evening.

Is there anywhere where the 0.6.9 DBus interface is actually documented? I
found the discussion on 0.7.2 proposals in the archives, but other than
trying to assess the source code I haven't found anything on 0.6.9 and I
haven't found any examples (even Python) that actually walk through
defining a net and selecting it. Nothing shows the actual messages that are
dissected to trigger the sequences. If you could lead me to something that
shows the sequence from the outside perspective that would help a lot.
Don Wilde
Waldo Health, Inc.
o: +11 512-219-0302
c: +11 512-749-6447

More information about the Hostap mailing list