The symptoms

• You try and listen to some songs and you get asked to authorise the computer. You do this but the next song owned by the same account asks you to authorise again.
• When you try and update applications for your iPhone you get an error -5000 on the downloads page
• As with the appstore, you get an error -5000 when podcasts are being downloaded.

The problem

This is such a trivial problem that it took a little while to work out what was going on. What’s happening is iTunes is unable to modify any of it’s files because of the permissions on the iTunes folders.

The solution

If your iTunes library is stored on an external hard drive (mine is), you can tell Finder to ignore permissions on the drive. To do this:

• Click on the drive in Finder
• Choose File -> Get Info (or press Command-I)
• In the bottom right of the information panel that appears, check the padlock is unlocked.
• If it isn’t unlocked, click it and enter your logon password
• Put a tick in the ignore ownership on this volume box

Alternatively, if your iTunes library is in your Music folder:

• Select your Music folder in Finder
• Choose File -> Get Info (or press Command-I)
• In the bottom right of the information panel that appears, check the padlock is unlocked.
• If it isn’t unlocked, click it and enter your logon password
• At the bottom of the information panel click on the gear icon
• Choose Apply to enclosed items – know this can take a while if your library is large

You should now find iTunes is working again.

Very simply you just need to give the website you are testing the email address username@computername.local. For example my username is matt and my computer is called tribble-mbp so my email would be matt@tribble-mbp.local. Alternatively you can instead use the computer name as localhost making the emailmatt@localhost.

Once you have the email set to your local machine you can use the built in mail program in MacOS.

Using command line mail

To use mail you need to fire up the Terminal.app (Finder > Go > Utilities > Terminal) and type mail. Most of the time you will be presented with a very boring response similar to this:

No mail for matt

If you do have mail you will instead see an interactive prompt similar to this:

>N  1 matt@tribble-mbp.loc  Tue Jun 16 10:40  14/493   "Test"
? 

This is your mailbox, and in my case shows I have 1 new message from matt@tribble-mbp.loc(al) which was sent on Tuesday June 16 at 10:20. The subject of the message is “Test”. This is helpful but how do I view the message?

Handy mail commands

Sending mail

The easiest way to send mail is to use the website you are developing but, should you wish to, you can use the command line tool. The best way to understand this is an example. We’re going to send an email to ourselves with the subject “Test” and the message “What’s up doc?”. To do this fire up Terminal.app and do the following:

• Enter mail matt@tribble-mbp.local replacing matt@tribble-mbp.local with the correct address for you.
• You will be prompted for the Subject: so enter Test and press enter/return.
• It now looks like nothing is happening but in reality mail is waiting for you to enter the body of the message. On the line that has just started type What's up doc? but do not press enter.
• To finish the message we have to ‘interrupt’ the program. We do this by holding down ctrl on the keyboard and pressing “ (sometimes you need to press it a couple of times).
• You should find the letters EOT appear and the program exits. If not, release any keys you have and try pressing ctrl and d again until it does.

When all is done the screen should look similar to this:

mail matt@tribble-mbp.local
Subject: Test
What's up doc?EOT

Great! you’ve sent your first mail. Lets take a look at it by entering mail in the terminal. You should find something similar to this:

>N  1 matt@tribble-mbp.loc  Tue Jun 16 10:44  14/486   "Test"
? 1
Message 1:
From matt@tribble-mbp.local  Tue Jun 16 10:44:10 2009
X-Original-To: matt@tribble-mbp.local
Delivered-To: matt@tribble-mbp.local
To: matt@tribble-mbp.local
Subject: Test
Date: Tue, 16 Jun 2009 10:44:10 +0100 (BST)
From: matt@tribble-mbp.local (Matt Harris)
What's up doc?

To delete the message type d1, replacing the 1 with the correct ID of the message.

Sending attachments

Sending attachments is again something you would probably want to do using your web application however it is quite straight forward. The method here does have it’s problems with encoding but works out of the box – as this is only for development and on the local machine it’s something I’m happy living with. To send an attachment use the following command:

uuencode /path/to/file.jpg filename-for-recipient.jpg | mail matt@tribble-mbp.local

This will send a single message with the @file.jpg@ attached to the message. Helpful but what about sending some text with the message as well? Well, this is achieved using echo and the | technique, like this:

(echo "My message test"; uuencode /path/to/file.jpg filename-for-recipient.jpg) | mail matt@localhost

Alternatively you can send a text file as message text using the cat command like this:

(cat mytextfile.txt; uuencode /path/to/file.jpg filename-for-recipient.jpg) | mail matt@localhost

Summary

Hopefully this has been useful for you. I found it partifularly helpful for debugging this. Knowing how mail works also means you can setup automated mailings from your MacOS to yourself, such as log files or alerts.

When you start using mail a lot, managing messages in the command line can become quite laboured so in my next post I’ll show you how to read the command line mail using Thunderbird.

What you will need

• The latest Android SDK
• At least version 1.5 of the Java Developer Kit. (type @java -version@ into your terminal or command prompt to find out the version you have).
• About 10 minutes of time

Setting Up

1. Follow the instructions on the Android SDK installation page making sure you follow the steps for adding the tools directory to your path. In my case I put the SDK into ~/Code/Android (where ~/ is my home directory).
2. Restart your command prompt or terminal and type android. You should be presented with the android help information. If not check you set the path correctly.

Target Platforms and Android Virtual Devices (AVDs)

It is possible to create emulators (AVDs) for each version (or target) of Android. To do this we need to check which hardware targets are available and what their target ID is. If you type android list targets into your terminal you will be presented with something similar to the code below.

Available Android targets:
id: 1
     Name: Android 1.1
     Type: Platform
     API level: 2
     Skins: HVGA (default), HVGA-L, HVGA-P, QVGA-L, QVGA-P
id: 2
     Name: Android 1.5
     Type: Platform
     API level: 3
     Skins: HVGA (default), HVGA-L, HVGA-P, QVGA-L, QVGA-P
id: 3
     Name: Google APIs
     Type: Add-On
     Vendor: Google Inc.
     Description: Android + Google APIs
     Based on Android 1.5 (API level 3)
     Libraries:
      * com.google.android.maps (maps.jar)
          API for Google Maps
     Skins: HVGA (default), HVGA-L, QVGA-P, HVGA-P, QVGA-L

From this information we can see the SDK supports Android 1.1 and 1.5 and their target IDs are 1 and 2 respectively. Knowing this information we can now create our Android Virtual Device. To create an AVD for Android 1.1 (target ID 1) with the name android1.1 we type:

android create avd -n android1.1 -t 1

Similarly for Android 1.5 we type:

android create avd -n android1.5 -t 2

A few things to know about AVDs. * When creating AVDs you can use any name you like. I choose android1.1 and android1.5 as they were easier to remember. * When prompted about hardware settings just press enter as there should be no need to customise the hardware when testing websites. * AVDs are created in a hidden folder in the root of your home directory. On Linux and MacOS this is ~/.android/avd. * You can find out what AVDs you have by typing android list avd in the terminal.

Running the Emulator

Having created the AVDs we can now run the emulator by typing emulator -avd avdname. This command is the one you will need to use from now on if you want to test websites in Android. For example, if we want to emulate Android 1.1 we would type:

emulator -avd android1.1

The first time you run the emulator you will be asked if you are happy for Emulator to send usage statistics to Google.

It can take about 10-30 seconds for the Android Emulator to boot up. During this time you will see the Android boot screen.

That’s it! To test your websites just load the Browser from the home screen. To enter a URL click Menu and then select Go or Go to URL depending on the version of Android you are running.

chmod -R 0755 directory
cd directory
find ./ -type f  -exec chmod 644 {} \;

The first command resets all files and folders to 755, allowing everyone to read and execute (or traverse for directories) and the owner to modify. The third command then resets just files to 644 which drops the allowed to execute flag, turning files from programs into files again.

I know previous parts of the site like the Flickr and Lifestream integration are missing – it’s ok, they will come back! – I’m just fixing up some of the new library I’m writing to manage them.

Anyway, let me know what you think of the site and more importantly if you have any problems viewing it.

A better way, which many of my professional colleagues and friends do, is to use MAMP or headdress. These two products provide a graphical interface to allow you to set up hosts on your computer and point the host to a folder on your computer. This is great except when you want to start developing with multiple languages, or wish to test SSL or external libraries. It also costs money as the virtual hosts ability is part of MAMP Pro, not the free MAMP edition.

So, although each of these products has their own strengths, is there a way to do what they do without using them. Luckily enough there is, and i’ll talk you through it now.

Assumptions

First off, i’m going to make some assumptions:

1. You’re using a Mac
2. You’re using Mac OS Leopard version 10.5.2 or higher
3. You know what the terminal is and
4. You know what the file paths are in MacOS
5. You’ve turned MAMP off and reset anything setup by Headdress

What you’ll need

• About 30 minutes of time (although you can do this in 2 minutes if you know your way around the shell)
• 4 files
  1. /etc/hosts
  2. /etc/apache2/extra/httpd-vhosts.conf
  3. /etc/apache2/users/USERNAME.conf (where username is the short username for your logon to your computer)
  4. /etc/apache2/httpd.conf
• A text editor. I use TextMate which means I can go into the terminal and type mate /path/to/file to edit the files. I’m going to use this notation throughout so remember to substitute mate with your favoured text editor.

Step 1: Allowing Overrides and Options

Open terminal and navigate to /etc/apache2/users and then list the contents of the folder.

cd /etc/apache2/users
ls -l

You will see a .conf file for each account on your computer. The one you want is the one which matches your short username for MacOS. In my case it is matt.conf

Edit this file in your favourite editor

mate /etc/apache2/users/matt.conf

By default this file will look like this:

<Directory "/Users/matt/Sites/">
    Options Indexes MultiViews
    AllowOverride None
    Order allow,deny
    Allow from all
</Directory>

You need to change the Options and AllowOverride lines to All, so the file should now look like:

<Directory "/Users/matt/Sites/">
    Options All
    AllowOverride All
    Order allow,deny
    Allow from all
</Directory>

Save the file.

Step 2: Configuring Apache

Open /etc/apache2/httpd.conf and make the following changes (the line numbers are correct for a fresh install but could be slightly different on your machine. The content remains the same though):

Save the file.

Step 3: Configuring your virtual hosts

Open /etc/apache2/extra/httpd-vhosts.conf. Everything after line 26 is an example of what needs to go in the file and can safely be removed. Alternatively you can comment the lines out with # if you prefer.

Here’s an example of what to put here, assuming:

• My website files are kept in /Users/matt/Sites/themattharris.com (remember this is case sensitive)
• I want to test the server with the DNS name www.themattharris.com

In this case I would put the following entry into the httpd-vhosts.conf file:

<VirtualHost *:80>
   DocumentRoot "/Users/matt/Sites/themattharris.com"
   ServerName www.themattharris.com
</VirtualHost *:80>

Step 4: Make the computer think the domain is on the localhost

Open /etc/hosts. You’ll see something like:

##
# Host Database
#
# localhost is used to configure the loopback interface
# when the system is booting.  Do not change this entry.
##
127.0.0.1   localhost
255.255.255.255 broadcasthost
::1             localhost
fe80::1%lo0 localhost

DO NOT change any of those lines, they’re critical to the workings of your computer. What you want to do is instead add lines to the end of the file, like the following for each of your virtual hosts:

127.0.0.1      www.themattharris.com

• 127.0.0.1 means my computer even when it’s not connected a network
• The domain name should match the one you used for the ServerName (or ServerAlias if you used that) in /etc/apache2/extra/httpd-vhosts.conf

h3. Step 5: Restart Apache

This can be done through System Preferences by turning Web Sharing off then on again. Alternatively, a quicker way is to use terminal and type:

sudo apachectl graceful

and enter you password when prompted.

All being well if you now point to the domain you setup in hosts you will see the web pages on your local machine. To add more virtual hosts just repeat step 3, 4 and 5.

Turning a virtual host off

When you want to visit the live site again you don’t need to change anything apart from edit /etc/hosts and place a # at the start of the line for the domain you want to turn off.

#127.0.0.1      www.themattharris.com

Troubleshooting

If you are having problems and the domains are not working open the apache error log (/private/var/log/apache2/error_log) and see what it says. Quite often you’ll find you’ve got the path in the wrong case and Apache couldn’t find the file.

Let me know how you do with this. In a couple of days I’ll show how to enable SSL on the virtual hosts and also how to compile the GD graphics library to work with MacOS.

October is Breast Cancer Awareness Month and so, to help raise awareness of breast cancer, Cindy drew me a pink version of my site which I have coded up for you all to see.

Now I’m not the only one doing this, lots of sites have gone Pink for October and you can find out more on that site. The BBC also has a useful health page if you are interested in learning more about Breast Cancer, the treatments, and also about early signs.

For those of you viewing this post after October 31st 2008 you can see how the site looked during the month in the screenshot on the left.

You can view the presentation “Adding a Timezone into your Upcoming Feed” on Google Docs

So you’ve installed Git on your remote server and now want to be able to keep connect to a remote Git repository. This post will go through setting up a connection to two popular repository hosting providers, Unfuddle and Github. I should stress these sites provide instructions on how to connect to their repositories and so what I talk about here is already on these sites; I just wanted to have a complete set of instructions available in one place.

What you will need

• SSH root access to your server (MediaTemple offer this though the Account Center
• An account on Unfuddle or Github
• An SSH Keypair setup on your server (if you’ve not done this yet see Generating an SSH Keypair
• About 15 minutes of time

Find out your SSH Key

Remote repositories use SSH Tunnels to transmit data between client and host. For this to work we need to register the SSH Key on our server with the remote repository. To find out what your SSH Key is you need to:

1. Log into your server using SSH. Usually this is ssh root@your-domain.com
2. Print the key to the terminal by entering the command : cat ~/.ssh/id_rsa.pub
3. Copy everything that was just output to the screen to a blank document or your clipboard. It will be something like ssh-rsa == root@

Registering our key with the remote repository

Now we have our SSH key we need to register it with our account on the remote repository. For Github this is on your account page under “SSH Public Keys”. On Unfuddle it’s on the “Personal Settings” page under “Public Keys”.

1. On Github choose to “add another public key”, whilst on Unfuddle choose “New Public Key…”
2. You’ll be asked for a title and a value. Enter a name in the title box which will identify this key clearly, for example “root on example.com”. In the value box paste the key pair you copied from the terminal window earlier.

These services now know who we are and we can authenticate with them. The next stage is to connect to the remote repositories.

Connecting to the remote repositories

1. First thing you’ll want to do is find the Git Clone address. On Unfuddle this is displayed under the name of the repository on the Repositories page. On Github the address is called the Public Clone URL and can be found on the Source page for the repository. It will look something like git@something.unfuddle.com:something/project.git or git@github.com/username/project.git
2. SSH into your server if your not in it any longer.
3. Navigate into the folder which will contain your clone of the Git repository.
4. Git cannot be cloned as a folder which already exists. If you want to clone the repository to a folder called “apples” you need to make sure the folder “apples” doesn’t exist already. In this case i’m going to clone my git repository into the folder httpdocs by using the command git clone httpdocs
5. If this is the first time you’ve connect to the repository you’ll likely be asked to accept the RSA key fingerprint. Check the host name and IP that’s displayed and if your happy answer yes
6. You’ll now be asked for the passphrase to your key. This is the phrase you entered when creating your SSH key pair before. This isn’t your SSH login password (unless you made them the same – which you really shouldn’t).
7. If you get errors here such as fatal: The remote end hung up unexpectedly make sure your password is correct and that you’ve registered your SSH key with your provider

You’ve now got a clone of the remote repository on your server. Now check the permissions are correct for your requirements. In my case I’ve cloned the repository into my web folder so I need to reset the permissions to those of my web account.

Now, whenever you want to update the copy of the code on your server just run git pull

In part 4 of this 4 part series I’ll talk about how I use Git to keep my production servers and easily publish updates.

The answer is remarkably simple and allows you to test any site hosted on your local machine with browsers installed on your virtual machines. In all of this I am assuming your running Mac OS X and Windows XP in the virtual machine. That being said, as long as your know your hosts IP address and the virtual machine is running Windows 95 upwards this will work. By host I mean the machine which you run VMWare or Parallels on.

What you need

• A virtual machine application such as VMWare or Parallels
• A website hosted on your local machine

What to do

With your virtual machine turned off, edit it’s settings to ensure it is running the Network card in “Bridged” mode. What this does is allow your virtual machine to appear on your network as if it was another computer, rather that the default NAT mode that hides the virtual machine behind the hosts IP address.

Next start up your virtual machine and whilst it’s loading, make a note of the hosts IP address. You can find your IP address in the network preferences or through the command ipconfig (Windows) or ifconfig (Linux and Mac).

With the virtual machine now loaded go to Start > Run and type notepad "c:\windows\system32\drivers\etc\hosts". This will load up the file Windows uses to alias IP addresses to host names.

The file is very well explained but in a nutshell, you put the IP address of the host at the beginning of the line, and the domain name that should point to it. So, lets assume my host IP is 10.0.1.97 and the domain I am developing is www.themattharris.com. In the hosts file I would enter a new line as

10.0.1.97       www.themattharris.com

Now lets assume i’m also developing another domain on the host called www.tokofu.com. All I need to do is add another line to the hosts file 

10.0.1.97       www.tokofu.com

Save the file and test in your virtual machine. The domains you entered into the Windows hosts file will have caused your web browser to request the site from your host and not the Internet.

We can also validate this by running two commands in the Windows Command Prompt. If we run nslookup the computer will lookup the domain name using a DNS server. If we run the command ping Windows instead uses the address it knows from the hosts file. You can see this by loading a command prompt (Start > Run and type cmd ) and entering nslookup your-domain.com and ping your-domain.com

To restore the default behaviour you edit the “c:\windows\system32\drivers\etc\hosts” file and either delete the lines you entered, or start the line with a # (this marks the line as a comment).