Linuxshoppen.dk has offered me to replace the Eee with a new one so I don't have to hate Asus support anymore \o/
So I'm going there tomorrow. Off course I have to see the new one actually working before I am happy but at least there seems to be light at the end of the tunnel.
- Open the PackageMaker in /Developer/Applications/Utilities
- Select "Assist me". I tried creating the package manually with "Single Package Project" but the PackageMaker crashed on me twice. It crashed when I was not even touching it but browsing on my Linux machine for stuff like "what does Remove .dsstore files "mean"?
- The "Assist" wizard only asked a few questions like the name of the package, company name, the path for the contents and the installation path.
- Then it built me a nice package that seems to work. I am installing it on top of an existing installation so I don't know for sure.
The package is not a file but a directory. It seems that the mpkg format is not a file either but also a directory. So I created a zip file. What is the convention for this on Mac? Do people send whole directories around?
Anyway next step is playing with the projector. Or perhaps getting the media for Mac OS X 10.5 and give it another try. (And see if I can still build packages
I wanted an Asus Eee 901 pc and I couldn't get my usual vendor to take it home for me. So I did what I ought to have learned not to do: Bought it from a vendor with whom I have no previous experience of support, and with whom I am not a big and well know customer.
I ought to know that falling in love online and just buying it like that was a bad idea. (Update: I got good service later, see later blog posts).
It was broken from the beginning and I already hate Asus support.
It goes into sleep mode randomly at inconvenient times. I tracked it down to being and acpi lid event that is sent every time I touch the lid, open or close it a little bit more, or touch the hinge. I tried with AC and no battery, with battery and no AC, I even tried reinserting the RAM because I found something on Asus forums about that having helped someone. No luck.
Then I stopped the acpi dameon (/etc/init.d/acpid stop) and that helped! But running it without ACPI was not really very satisfactory so I looked in /var/log/acpi and saw that the last event I had had was a lid event. Ahh time for some debugging!
cat /proc/acpi/event > button/lid LID 00000080 00000036 > button/lid LID 00000080 00000037 > button/lid LID 00000080 00000038 > button/lid LID 00000080 00000039 > button/lid LID 00000080 0000003a
etc etc every time I moved the hinge or came anywhere near it. I tested the power button, the sleep button and actually closing the lid : One correct acpi event for each action. But touching and fiddling with the lid and hinge gave sometimes 3 or 5 acpi events.
So I wrote to linuxshoppen.dk late last night.
Half an hour ago I got the answer, no they never heard of it before, call Asus hotline.
Thank you very much for ruining my day. Calling a hotline is guaranteed to take my blood pressure to a level where I get no work done whatsoever for several hours after.
But I checked on the web that they are only open until 5 pm so I had to call now.
So I called Asus hotline. They actually didn't ask me for a serial number (yet). They didn't even really want to hear any details about the error. I probably go in the stupid blond category by default (or perhaps they have no other categories)?
Anyway I was told to do a recover install or whatever they call it, because it might be a software problem. I hardly think so.
He didn't think so either, but said that if I send it in for repair and it is not a hardware problem, I shall be billed 1000 kr. Thank you it did the bloody thing from the moment I bought it! I didn't install the software, Asus did!
I don't mind at all doing the software recovery install. It is a fine procedure to have people do that first. It probably fixes a lot of problems without people having to send the laptop in.
I do mind sending it it for repair with the threat over my head that it they cannot find the hardware problem they will bill me 1000 kr for it. I want a guarantee that if a software reinstall did not solve the problem I shall not be billed for the repair.
So now I am going to waste a lot of time and emotion on something that should have been a handy tool and cool toy.
Why is everybody else having fun with their and mine has to be broken? I feel a bit like throwing it in the bin and just getting another Thinkpad, but that is just me overreacting because I am disappointed.
More to follow.
Took some time and some frustrations because I am such a young person My first Internet connection was a LAN connection, I never really owned a modem before so I don't speak AT commands fluently (yet).
There are several ways to do this. In the end I guess they all do the same, you just have to decide how manually you want to do it.
I think the modem is a HUAWEI E169 but it seems to behave exactly the same way as a E220, I haven't figured the difference. Someone else might elaborate on that. It works as a usb storage device at the same time if you add an sd card to it so don't worry when you see the usb storage module being loaded.
I needed to know what my device name was and that changed a lot when I initially experimented. I ended up with /dev/ttyUSB0 though it looks like 3 devices to dmesg:
dmesg |grep ttyUSB usb 2-2: GSM modem (1-port) converter now attached to ttyUSB0 usb 2-2: GSM modem (1-port) converter now attached to ttyUSB1 usb 2-2: GSM modem (1-port) converter now attached to ttyUSB2
That seems to be OK, there should be 3. This page http://wiki.debian.org/Modem/3G/Vodafone says a little about finding the device.
The most gui-ish way is network manager (didn't work for me).
- Guide for Asus eee 900: http://www.linuxshoppen.dk/3guide.php (in Danish).
- Network manager 3 option: http://www.evil.ie/bleh/?p=113 has a link for a /usr/share/hal/fdi/information/20thirdparty/10-3g-modem.fdi file so network manager recognizes a 3 connection as an option (under VPN). New versions of network manager might include this.
Anyway Network Manager in FC8 seems to be very broken and in FC9 I couldn't save my settings so I gave up.
Next options was to go ppp directly, as this config file for FreeBSD shows (kindly provided by phk):
http://archive.bsd-dk.dk/mail/bsd-dk/2008/Sep/0016.html (in Danish). I don't think this one maps directly onto Linux but to get some inspiration about syntax, see http://forums.whirlpool.net.au/forum-replies-archive.cfm/824630.html . I think these two together would have got me up'n' running if I hadn't got the third solution to work:
My solution was wvdial. Several config files can be found online, especially this page has configs for many countries: http://ph.ubuntuforums.com/showthread.php?p=5416802. Thank you to the author of that one! http://www.evil.ie/3/wvdial.conf also has a wvdial config.
However, none of those configs were exactly right, though close, and I needed a bit more.
So this is my wvdial.conf:
[Dialer Defaults] Init2 = ATQ0 V1 E1 S0=0 &C1 &D2 +FCLASS=0 Modem Type = Analog Modem ; Phone = <Target Phone Number> ISDN = 0 ; Username = <Your Login Name> Init1 = ATZ ; Password = <Your Password> Modem = /dev/ttyUSB0\nBaud = 9600 [Dialer pin] Modem = /dev/ttyUSB0 Init1 = AT+CPIN=**** <- this is really my pin code [Dialer hsdpa] New PPPD = yes Phone = *99# Username = irrelevant Password = irrelevant Stupid Mode = 1 Dial Command = ATDT Modem = /dev/ttyUSB0 Baud = 460800 Init2 = ATZ Init3 = ATE0V1&D2&C1S0=0+IFC=2,2 ISDN = 0 Modem Type = Analog Modem Init5 = AT+CGDCONT=1,"IP","data.tre.dk"
On top of having a wvdial file I also have a /etc/ppp/options file to include some of the options in the freebsd example:
lock lcp-max-configure 10 mru 1440 nobsdcomp nodeflate nopredictor1 novjccomp nomppe usepeerdns
lcp-max-configure was something I found while debugging, you can set it higher to wait longer before giving up on something called "LCP: timeout sending Config-Requests" (which I only got while trying on /dev/ttyUSB2 instead of ttyUSB0).
usepeerdns gives me a /var/run/ppp/resolv.conf with the "3"
name servers in it.
They are probably called something else in phone terminology, but I call them server names. In Denmark "3" has 3 different options:
- bredband.tre.dk - not firewalled, public IP
- vip.tre.dk - partly firewalled, public IP
- data.tre.dk - firewalled, private IP
according to http://www.servernet.dk/?m=200708. Roaming should work better on data.tre.dk where you get 10.x.x.x addresses. On the other 2 you get 94.x.x.x and 80.x.x.x addresses (I did yesterday when I tested, at least).
The firewalling part seems to be only ingoing, at least I can access port 22 outgoing and that is all I need, except from off course port 80, udp port 53 etc. but all customers need that so that must be open.
My start script
#!/bin/sh sudo /etc/init.d/network stop sudo wvdial pin >/dev/null 2>&1' sleep 5 sudo wvdial hsdpa
- the redirect to /dev/null is so people on the train etc. cannot see my pin code on the screen. I don't know if the sleep command is necessary, really. I call it "3" and I just run it from a terminal, and ctrl+c it when I finish using the connection. There might be a nicer way...
So, this better work on the train tomorrow! Then it will be worth all the trouble.
I upgraded my laptop to fc9 because I wanted to get 3 mobile broadband to run via network manager. That did not work out. Network Manager is broken I think, I got it to run more or less with wvdial. Blog post on that later, it doesn't quite work right yet.
But after an upgrade some stuff has always gone awry. My wireless for example. And I need it tomorrow ...
So I had to figure it out today.
What has happened is that my ipw3945 intel card does no longer need a commercial driver called ipw3945 and a corresponding daemon called ipw3945d, instead it uses a built in driver called iwl3945.
That sounded fine with me, except it didn't work. At least not with wpa2 eduroam it didn't.
So I googled evil error messages like
ioctl[SIOCSIWAUTH]: Operation not supported
for a long time. But that was not the point.
It seems to be ok that these error messages are there, since the driver just don't support these features yet.
But it didn't work.
Until I found this page: http://wiki.debian.org/iwlwifi telling me not only to uninstall all the ipw3945 stuff including the ipw3945d, but also to remove a line in /etc/udev/rules.d/z25_persistent-net.rules concerning ipw3945. On Fedora that line is in 70-persistent-net.rules.
All of a sudden it is running perfectly again, in spite if the error message. Wee \o/
I need to start wpa as a daemon which I do with /etc/init.d/wpa_supplicant start. After that I say /etc/init.d/network restart but perhaps an ifup of the eth1 interface could do it.
My /etc/sysconfig/wpa_supplicant has not changed:
The error messages I get when running wpa manually with
wpa_supplicant -ieth1, -Dwext -c/etc/wpa_supplican /wpa_supplicant.conf
ioctl[SIOCGIFFLAGS]: No such device Could not get interface 'eth1,' flags ioctl[SIOCSIWPMKSA]: No such device ioctl[SIOCSIWMODE]: No such device Could not configure driver to use managed mode ioctl[SIOCGIWRANGE]: No such device ioctl[SIOCGIFINDEX]: No such device ioctl[SIOCSIWENCODEEXT]: No such device ioctl[SIOCSIWENCODE]: No such device ioctl[SIOCSIWENCODEEXT]: No such device ioctl[SIOCSIWENCODE]: No such device ioctl[SIOCSIWENCODEEXT]: No such device ioctl[SIOCSIWENCODE]: No such device ioctl[SIOCSIWENCODEEXT]: No such device ioctl[SIOCSIWENCODE]: No such device ioctl[SIOCSIWAUTH]: No such device WEXT auth param 7 value 0x0 - Failed to disable WPA in the driver. ioctl[SIOCSIWAUTH]: No such device WEXT auth param 5 value 0x0 - ioctl[SIOCSIWAUTH]: No such device WEXT auth param 4 value 0x0 - ioctl[SIOCSIWAP]: No such device ioctl[SIOCGIFFLAGS]: No such device
Update My device is no longer eth1 but wlan0.
It was eth1 the first time but that was probably a leftover from the ipw3945 driver having been loaded. After reboot without it I had wlan0 instead.