Although a total newbie to MAC OS X, I thought I would post my experiences getting my iMac G3 to access the Internet by way of my Tecom BT3021 Bluetooth Access point. I am running MAC OS X 10.4.3 Tiger, and use a third party USB Bluetooth dongle on the iMac -- not one of the officially supported D-Link ones, but it works OK nonetheless.

Having waded through several forum threads which mentioned that the former Bluetooth Serial Utility had been pulled from this release of OS X, I managed to download this utility and found that it was invaluable in allowing me to set up the outgoing serial port to use the LAN Access via PPP profile provided by the Access Point. Having done all that, I found that the connection would not complete due to LCP errors, namely thisone:

LCP: timeout sending Config-Requests  : (date and time)
  Connection terminated.

Deciding that the Access Point, being a fairly dumb device (no disrespect intended), would need to be involved in as little config negotiation as possible, I set about editing the /etc/ppp/options file. Presto! I was able to get this to work. Here is my edited file.

As I said at the outset, I am very new to OS X, and also to Unix generally, so the above may not be entirely optimal -- for example, the mru and mtu settings could be better tuned for this type of connection. However, I thought I would share my first "success" with OS X with the wider community. Comments welcome.

[robg adds: I believe all of the functionality of the old Bluetooth Serial Utility has been replaced by the Devices tab of the Bluetooth System Preferences panel (which now includes a Serial Ports button). However, I didn't use the prior program extensively, so that may not be correct.]

I had an old Linksys WRT54GS router that I wasn't using, and because of the way my house is set up, I needed to put my AirPort Extreme Base Station somewhat far from my home office. This was a problem for two reasons: first, I had a relatively weak signal in my office, and second, because I have a network attached storage system and an Ethernet-based printer in my office that I'd wanted to keep there. I knew that the AirPort Express would accommodate this setup, but since I already had the Linksys -- which doesn't have an audio out or a USB port, like the Express, but which does have four ethernet ports -- I thought it was worth trying.

There are tons of tutorials online for using the Linksys as a Base Station, with the AirPort Express as the repeater/secondary access point. (The protocol for this is called WDS. Apple's routers support it; the Linksys does as well. Others may, as well -- but Netgear doesn't, and apparently can't.) I couldn't find any decent tutorials for reversing the process, with the Linksys as slave to an AirPort.

This is a little complicated, and might require some trial and error, but here are the basic steps.

[robg adds: Somewhat obviously, I haven't tested this one. The following instructions are detailed, and may be overly so for experienced users. However, those who are trying to figure it out for the first time may appreciate the detail. Hopefully I didn't introduce any errors while formatting the article.]

I pretty much always have my home file server mounted via samba (automount/dynamic) on my laptop. The problem is that samba and other network shares usually don't play nice after a sleep, causing me to manually remount the shares. This is especially the case when my ip changes. My solution is to use the sleepwatcher daemon (download) to run commands just before sleep and after wake.

By unmounting your network shares just before sleep, your system won't have to deal with lost connections when it wakes up. Additionally, when your system wakes up, the sleepwatcher daemon will run your wakeup script, which can remount your network shares. sleepwatcher daemon Here are the steps:

1. Install sleepwatcher.
   
   
2. Make a shell script in your home directory named .sleep. I just unmount my network mounts in this script. Make sure you chmod it to at least 700; 755 is OK, too. Here's the script:

   logger -t sleepwatcher "unmounting serv"
   umount /Network/Servers/servh/data

3. Make a shell script in your home directory named .wakup. Since this script might run before the network comes back up, I use a trick to pipe the command to a backgrounded bash prompt with a sleep 10 command. chmod this one to 700 or 755 as well:

   echo 'sleep 10
   logger -t sleepwatcher "remounting serv"
   ls /Network/Servers/servh/data' | /bin/sh&

The logger lines help you see what's going in system.log. Since my mount is dynamic, a simple ls is enough to tell automount to mount it.

I often access my G5 and my home network at large from the office. The problem is my bad connection, NAT, and not having a static IP. It's usually good enough that I only have to check my WAN IP address once a day. The problem is I'm going on vacation for a while, and would like to be able to 'check in.' So, between my abysmal scripting and Automator skills, I think I've got something worked out.

So, the first thing is a bash script that goes out to the web and gets the IP address of the machine it's running on. If things worked, it will print a message saying:

Here's the current IP Information for DuoBook

XX.XX.XX.XX
As far as I can tell everything ran fine.

Next, I compiled an Automator workflow (download) into an application to email me that output, and clean up the temp files. It's running in a cron job.

It seems to be running fine, although, like I said my scripting / Automator skills are terrible, so it could break. If anyone sees anything wrong, could you let me know before I leave for that vacation. You'll have to tweak the workflow, as it's got my email built in, and it relies on knowing where the script is.

I listen to music quite often on my PowerBook whilst studying, but, like most people, I find hearing it through the built-in speakers sacrilege. Aside from setting up my desk next to a stereo, the first solution that comes to mind is Apple's AirPort Express with Airtunes; this is essentially a wireless router that has a connection to your stereo, and has the ability to play audio from iTunes. This all comes at a price, however.

The following guide will show you how to live the champagne lifestyle on a cask-wine budget (at least with respect to your audio setup).

The solution

Thanks to the efforts of hard working and hospitable geeks, a free, open-source solution is at hand -- and has been since about 1998. Its name is the Enlightenment Sound Daemon (esd). The Enlightenment Sound Daemon is essentially a network-transparent audio protocol, in many ways is analagous to X11 with graphics (if you don't know what X11 is, then you're probably not geeky enough for this tutorial -- buy an AirPort Express).

Just to post a follow-up for people who were trying to connect to a standard install of WinServer 2003, and were unable to do so, due to error -5000. I was getting errors every time, regardless of the user settings.

The referenced hint above requires you to downgrade the security settings on the server. Instead, I changed one line in the file /etc/smb.conf/. Change the line that reads:

  client ntlmv2 auth = no

to this:

  client ntlmv2 auth = yes

After making this change, it works again. No need to restart or anything. This may not work for everyone, but it does solve the issue for me.

The iDisk feature of .Mac can be very useful, however it seems to be plauged by slow speeds (I find it impossible to use without the local iDisk turned on) and sync issues (automatic sync causes corrupted files in some apps). Even with auto-syncing off, open/save dialogs and even opening Finder windows can cause a spinning rainbow cursor; sometimes for over a minute.

The problem lies in the fact that the local iDisk (stored in /Volumes/iDisk) has symbolic links (aliases) to folders on the remote iDisk (/Volumes/your_mac.com_name). When OS X lists the local iDisk's contents, info for these folders must be loaded from the remote iDisk, which can take some time. The Open and Save dialogs are particularly bad if they default to a location on the local iDisk, because the column view that displays the files and folders includes the alias folders at the root of the local iDisk.

As some people have noted, the problem gets better after the first incident. This is because OS X is kind enough to cache the remote folder info. However, I have found that the problem can persist and present itself in various ways.

With the new child in the house, I've lately found myself working in many locations at many odd hours of the night. The end result is that my 12" PowerBook is getting a lot of exercise lately, including a lot of hint writing and editing. (And it's also making me yearn for a 15" Core Duo machine; as much as I love travelling with my 12" PB, it's not the ideal machine from which to work remotely within the house. The lack of pixels leads to too much reliance on Exposé to help get things done.)

My Dual G5 desktop runs a local copy of Geeklog, so I do all the previewing and editing of stories from its database on the G5. When working on the PB, I enter the G5's IP address to load the submitted stories, preview my edits, and then publish to the G5. (Later on, I dump the SQL data files from the G5 and upload to the web server to publish the hints.)

I was doing this the other evening when the G5 suddenly stopped serving pages. So I put down the PB, walked over to the den, and found that the G5 had decided it was time for its pre-set "sleep if idle for 30 minutes" timeout. But clearly the machine wasn't idle -- it was serving web pages to the PowerBook. I guess, though, from the system's perspective, this didn't show as activity as far as the sleep schedule was concerned (no mouse movement or screen updates).

I didn't really want to disable sleep on the G5, so after a bit of expermentation, I found the solution: to keep a networked machine awake, you simply have to connect to it via file sharing. On the PowerBook, I mounted my home directory from the G5, and went back to work. Hours later, the G5 was still awake. I unmounted the drive when I was done, and the G5 fell asleep again as scheduled.

I assume there are other tricks to keeping a "non active" networked machine awake (setting a cron task to do something every so many minutes?), but this one worked for me.

I'm currently a student at Davidson College. The school currently protects its network with a program called Cisco Clean Access. A compulsory client exists for XP machines that enforces all sorts of local policies (like running a virus scanner) before automatically providing the machine access to the campus network after the user provides their login name and password.

On non-XP machines, a user has to provide their credentials via a web-based form before access to the network is granted. I have a MacBook Pro, and I've gotten really sick of having to re-login to the campus network (both Ethernet and wifi) via this webform every time my computer goes to sleep or I change locations. So, I've managed to write a bash script that uses curl to log me in through the web forms.

Thanks to some amazing work by macrumors forum member Wombert, I've also found a way to have it activate automatically, any time the campus wifi network SSID is detected. This automation avoids heavy-handed cron jobs by leveraging two OS X features, configd and Kicker.xml. (There's a previous hint here that relies on AppleScript and iCal scheduling.)

Now the script and Kicker.xml transparently log me onto our campus network, and hence the internet, any time my computer senses a preset group of SSIDs. The curl commands are specific to Cisco Clean Access, but they could easily be adapted to other environments where users must login through web forms, making this script potentially broadly useful.

One downside to the script right now is that it stores the username and password in plaintext. Perhaps someone could make this part more secure with more sophisticated code. Instructions are included in the comments of the script. To use the script, just copy and paste into a .sh file, and make it executable. You can also see it all marked up at pastebin.com.

[robg adds: I have not tested this one...]

After using OSXvnc for a while I installed Apple Remote Desktop client 2.1 because I have two screens. Chicken of the VNC to ARD works great (though slowly) as long as I do not choose 8-bit mode and remember to send a password. That is not the problem. The problem is that I had OSXvnc working over ssh, but ARD added a firewall entry for itself that opens ports 5900 and 3283 every time I start it. This was a general problem for me; I did not like how enabling a service would open ports for it in the firewall, since I forward everything I can over ssh. In 10.3, I had a startup item that would configure ipfw itself, but you are supposed to use launchd in 10.4. So I decided I would figure out how to not have ports open automatically when I start a service.

[kirkmc adds: I had a bit of a problem testing this hint (well, at least verifying the steps; I didn't fully test it). The author mentions expanding "udport" in the plist file in Property List Editor, but I didn't find that key. I tried activating ARD VNC access, and setting a user and password, but the key still didn't show up.

By iChat, I asked Rob if he had it, and indeed he does. I don't know why there is a difference, so if you don't see it, well, I don't know what to suggest. Since this ARD access is a system function - you don't need to have a copy of ARD installed - there's no reason why one Tiger system would have it and another wouldn't.]