haha.. can't believe it took me this long to catch this oversight. I used glade to design the UI for configuring the connection, and it just defaults to a top-level window. I'm loading the glade ui file as a resource and I never really closely exampled the XML that glade-ui designer output. Oops.
Gonna need to completely re-work my UI code to fix this. boo.... That's not trivial undertaking (for me). I'll start on it in the next few weeks, now that I know what the problem is. Thanks for the insight! On another note, and somewhat a shot in the dark, I hit a bug yesterday while testing my code that I never had before. I wanted to manually verify an IPsec connection before I tried my gui again, and I got an error messge "error inserting padlock, no such device" blahblahblah. I only get it when I try to start a connection that uses AES encryption. I'm using openswan as my IPsec implementation. Anyone else hit this? I'm using a dell laptop, so I don't think this is related to the VIA Bios issue I've read about on the 'net. Any help greatly appreciated. I'm not down with 3DES and all my test connections use AES encyption, so testing has become somewhat impossible until I find a fix for this. I know I can upgrade, but Fedora6 is rock-solid stable and I love my beryl-setup with all that eye-candy. Can't get it to work in FC7. All help appreciated. steve. Turbo Fredriksson wrote: > Quoting steve hillier <[EMAIL PROTECTED]>: > > >> So, first you need to get nm >> "aware" of the nm-openswan vpn service (see above). >> > > Being a Debian GNU/Linux developer, the first thing I did (after trying > it 'your way' :) was to create a .deb package. All this works perfect. > Got network manager to see the new service etc... > > >> You'll know it's >> working, when you can add the "IPsec" type of vpn connection through >> the vpn manager interface. Selecting this will start a wizard which >> eventually calls out to my shared library to configure the connection >> details. >> > > This is where it breaks. SOMETIMES (especially if I'm running with > strace or gdb), i can briefly see a window pop up. But this is where > it segfaults... > > >> If you're debugging the shared library, it's at this point, >> you're debugger will go active (when nm tries to load my library and >> call the factory method to instantiate the configuration window). This >> is where the library tries to build a GUI and crashes. >> > > Right.. > > >> I Just recently I realized that I may be creating my config window >> entirely wrong. I'm buidling a top-level window when I should be >> adding elements to the wizard window (I think). >> > > Seems resonable, because nm say: > > (nm-vpn-properties:20827): Gtk-WARNING **: Can't set a parent on a toplevel > widget > > Unfortunatly, gdb doesn't catch the segfault so I can't get a back > trace... And I can't get anjuta to recognize the project. > > And I'm as green as you when it comes to coding GUI's with GTK... > > _______________________________________________ NetworkManager-list mailing list NetworkManager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
