#!/usr/bin/env bash
 
if [[ $# -eq 0 ]]; then
	dir=`pwd`
else
	dir=$1
fi
 
/usr/bin/osascript \
-e "tell app \"System Events\" to tell process \"Terminal\" to keystroke \"t\" using command down" \
-e "tell app \"Terminal\" to do script \"cd $dir\" in front window" \
-e "tell app \"System Events\" to tell process \"Terminal\" to keystroke \"l\" using control down"

I’ve put the shell script in “/usr/local/bin” as this is in my $PATH. I’ve named the script “tab” so whenever in Terminal I can enter “tab” to open a new tab with the current working directory ready and waiting. Very useful when working with Rails and multiple shells.

function clear_field(el, handle_blur) {
  $(el).value_modified = false;
 
  $(el).onfocus = function() {
    if ($(el).value == $(el).defaultValue && $(el).value_modified == false) {
      $(el).value = '';
      $(el).value_modified = true;
    };
  }
  if (handle_blur !== false) {
    $(el).onblur = function() {
      if ($(el).value == '') {
        $(el).value = $(el).defaultValue;
      };
    }
  };
}
 
window.onload = function() {
   clear_field($("login"));
   clear_field($("password", false));
}

The “clear_field” function handles focus and can handle blur events. When a field receives focus for the first time the “value_modified” boolean variable is set to true and the value of the field is set to an empty string. Subsequent focus events will no longer empty the field as we do not want to interfere with the user’s interaction with our field.

If we supply a “false” argument after our element the blur event will not be applied. If this is the case the field will never be repopulated. Handy for password fields, as is indicated in the example above.

If the blur event is applied the field will be repopulated with it’s defaultValue if the field is empty when it loses focus.

I’ve uploaded a demonstration of the use of this code. You can find it at http://static.jamesconroyfinn.com/clear_field/.

You’ll probably find your mod_default module is already installed and available with just a couple of lines of configuration.

I’ve enabled mod_deflate globally across all my domains and hence have added my configuration to my main httpd.conf file. With most flavours of UNIX it’s located in /etc/httpd. If not consult the documentation for your distro. Should you only wish to enable for a single domain you’d need to add the AddOutputFilterByType and BrowserMatch rules below to the VirtualHost section in your configuration.

These instructions are tailored for my setup (a (dv) 3.5 with MediaTemple.) Your host may put your configuration files elsewhere. You will need root access via SSH to your web server to complete these modifications. If you aren’t comfortable in the command line best you ask your host very nicely to do this for you.

First off we need to make sure mod_deflate is being loaded by Apache. Fly in to the command-line and use…

# grep 'mod_deflate' /etc/httpd/conf/httpd.conf

…to search for line that will load the mod_deflate module. If it’s commented out with a leading “#” then we need to remove the hash sign in order for mod_deflate to be loaded when we start up Apache.

I personally used vi to do this but you could use perl from the command line if you wished to. Remember to back up your files regularly and do not make modifications to httpd.conf without a stable copy you can revert to should things go wrong!

# perl -pi -e 's/# LoadModule mod_deflate/LoadModule mod_deflate/g' /etc/httpd/conf/httpd.conf

This command searches for an occurrence of “# LoadModule mod_deflate” in your configuration file and replaces if with “LoadModule mod_deflate” essentially just stripping the hash sign from the line.

Once we have an entry in httpd.conf for loading mod_deflate we need to add configuration settings for mod_deflate. Again, I used vi for this.

# vi /etc/httpd/conf/httpd.conf

In you configuration file you need to add…

#
# Deflate output configuration
#
AddOutputFilterByType DEFLATE text/html text/plain text/xml text/css
BrowserMatch ^Mozilla/4 gzip-only-text/html
BrowserMatch ^Mozilla/4\.0[678] no-gzip
BrowserMatch \bMSIE !no-gzip !gzip-only-text/html

I put these lines in with some of the shtml configuration options near the top of my file. Just make sure you don’t nest the options inside a particular module’s configuration section or within a VirtualHost definition.

I used grep to pull the lines out of my http.conf file below. The -B flag pulls a number of lines before a match is found and the -A pulls lines after. The command below will find any occurrence of “Deflate” in httpd.conf and print out one line before and six lines after.

# grep -B1 -A6 'Deflate' /etc/httpd/conf/httpd.conf
#
# Deflate output configuration
#
AddOutputFilterByType DEFLATE text/html text/plain text/xml text/css
BrowserMatch ^Mozilla/4 gzip-only-text/html
BrowserMatch ^Mozilla/4\.0[678] no-gzip
BrowserMatch \bMSIE !no-gzip !gzip-only-text/html

With mod_deflate enabled and our configuration settings added to compress HTML, plain, XML and css files we’re almost ready to restart Apache. Before doing so we always need to check syntax in our config files using…

# apachectl -t

If we get a “Syntax OK” message back we’re good to initiate a graceful restart.

# apachectl graceful

Now we should find Apache pushes files out using gzip compression, hence reducing page loading times and making the end user experience that bit more pleasurable.

There is a similar article on howtoforge regarding setting up mod_deflate on a Debian system. Check it out at HowToForge.com.

I found the plist ruby gem to be somewhat unreliable but think I might have found a way to temporarily provide all the functionality required by piggy backing on plutil. The email I sent to Patrick is below as I don’t want to have to type up the same waffle twice.

Hi Patrick,

I’m working on sorting the network connect plist so that System Preferences and the VPN menu will display connections sorted alphabetically.

Initially parsing…

File.expand_path("~/Library/Preferences/ByHost/com.apple.networkConnect.#{mac_addr}.plist")

…returned the old unknown elements runtime error. Regrettably I can’t send the entire contents of my preference file as it contains some sensitive information even without passwords.

It did make me think about an alternative approach however, which allowed me to parse the plist successfully. I used plutil to convert the plist file to an xml1 format…

plutil -convert xml1 -o /tmp/plist.xml com.apple.network…

Then by parsing the xml I had full access to the file. Of course, one could use Ruby’s built-in temporary file functionality rather than writing to the /tmp directory as I did.

This could theoretically be included as a fall back should an unknown element be encountered.

Thanks for the effort you’ve put in to developing this gem. It’s a handy piece of kit!

Kind regards,
James

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

#!/usr/bin/env ruby
 
begin
  require "rubygems"
  require "plist"
rescue LoadError
end
 
# def print_usage
#   puts "Usage: #{File.basename(__FILE__)} [file_path]"
#   exit
# end
 
if ARGV.length == 1
  file_path = File.expand_path(ARGV[0])
else
  mac_addr = `ifconfig en0 ether`.split("n")[1].gsub(':', '').match(/[da-f]{12}/)
  puts mac_addr if $VERBOSE == 1
 
  file_path = File.expand_path("~/Library/Preferences/ByHost/com.apple.networkConnect.#{mac_addr}.plist")
end
 
if File.exists?(file_path)
  # We don't want to worry about overwriting files
  uniq_file = "/tmp/plist.#{(rand * 10000).floor}.xml"
 
  # Thank you Apple, I can never stay mad at you!
  %x(plutil -convert xml1 -o '#{uniq_file}' '#{file_path}')
 
  # Parse our valid xml file
  plist = Plist.parse_xml(File.expand_path(uniq_file))
 
  # This is everything that comes out
  puts plist.inspect
 
  # Clean up, just for fun!
  FileUtils.rm(uniq_file)
else
  raise ArgumentError, "Invalid path specified "#{file_path}""
end

Don’t use this code AS IS! It’s a proof of concept. It doesn’t work and the really hard graft is yet to come. This just shows it can be done!

The problem with the (mt) article is it’s outdated. The URI they provide with the initial rpm command is non-existent. It produces a lovely little 404 error, which rpm reports as follows…

# rpm -Uvh http://mirror.centos.org/centos/5.1/os/i386/CentOS/yum-3.0.5-1.el5.centos.5.noarch.rpm
 
Retrieving http://mirror.centos.org/centos/5.1/os/i386/CentOS/yum-3.0.5-1.el5.centos.5.noarch.rpm
error: skipping http://mirror.centos.org/centos/5.1/os/i386/CentOS/yum-3.0.5-1.el5.centos.5.noarch.rpm - transfer failed - Unknown or unexpected error
warning: u 0x807d980 ctrl 0x807e148 nrefs != 0 (mirror.centos.org http)

First of all we need the “yum-metadata-parser.” If you don’t install this dependency the yum install will fail.

# rpm -Uvh http://mirror.centos.org/centos/5/os/i386/CentOS/yum-metadata-parser-1.1.2-2.el5.i386.rpm

Now we can install yum itself.

# rpm -Uvh http://mirror.centos.org/centos/5/os/i386/CentOS/yum-3.2.8-9.el5.centos.1.noarch.rpm

This link may no longer be accurate in the future as the rpm package includes a version number. If CentOS decide to no longer mirror the download we’ll need to track down the latest version and rpm that instead.

Once this is all installed we can update yum using “yum update.” When I ran the update there were roughly 142 packages to update totalling around 245 MB of actual data. It was done pretty sharpish thanks to (mt)’s huge pipe!

The git package is available via EPEL repository so we need to run another rpm command in order to search this repository when using yum.

# rpm -Uvh http://download.fedora.redhat.com/pub/epel/5/i386/epel-release-5-2.noarch.rpm

With yum up to date and access to the EPEL repository we can finally install git.

yum install git

When I first downloaded it I was impressed but not exactly blown away as the snippets don’t really leverage the true power of TextMate’s bundle system. I had to get to work to make things a lot smoother.

/*
	${1:Working Title}
	Updated:	${2:`Date +'%a'`} ${3:`date +'%d/%m/%Y'`} ${4:`date +'%H:%M:%S %Z (%z)'`}
	Author:		${5:`dscl . -read /Users/$USER RealName | awk 'NR==2 { sub(/^[ \t]+/, ""); print }'`}
	----------------------------------------------------*/$0

I modified the “comment: updated” snippet to automatically populate the Updated timestamp for us and added a dscl command to get the RealName associated with the current user’s shortname.

The date timestamp will appear in the format “Mon 23/04/1984 12:01:21 GMT (+0000).” I’m using this format as I’m British and it’s makes sense to me. If you’re American you could switch the “%d” and “%m” around to get “MM/DD/YYYY.” Alternatively you could just replace “date +’%d/%m/%Y’” with “date +’%D’” to have a date formatted according to your computer’s default date format. For more all the possible options available when formatting dates using the date command check out WikiPedia’s date article.

I also amended the “comment: declare” snippet similarly so my RealName would be added automatically to the snippet. To do this you only need to replace “${4:Your Name}” on the “Author:” line with…

${4:`dscl . -read /Users/$USER RealName | awk 'NR==2 { sub(/^[ \t]+/, ""); print }'`}

Or just replace the whole snippet with my amended version below.

/*   
Theme Name: ${1:Name}
Theme URI: ${2:URL}
Description: ${3:a-brief-description}
Author: ${4:`dscl . -read /Users/$USER RealName | awk 'NR==2 { sub(/^[ \t]+/, ""); print }'`}
Author URI: ${5:your-URI}
Template: ${6:use-this-to-define-a-parent-theme--optional}
Version: ${7:a-number--optional}
.
${8:General comments/License Statement if any.}
.
*/$0

Ok, so maybe I don’t love my iPhone enough to engage in autoerotica but with the new apps I have access to I do find it much more productive!

The first couple of apps I downloaded were Apple’s Remote, Facebook and Twitterrific.

Remote is brilliant. Imagine listening to your music on your iPhone with all your glorious album artwork displayed on the big bright screen, playing, pausing and scrubbing away. Now imagine the music that’s playing is actually coming out of your £2000 surround sound stereo system and your 8 GB iPhone has access to 500 GB of music. All you need is iTunes on the same wireless network as your iPhone and you can pair away. The integration with iTunes feels solid and professionally done which really helps!

If you use Facebook then the Facebook app is a must have. The initial release had all the most basic features of Facebook without all the adverts and tacky third-party apps. Since 1.1 the Facebook has seen a few improvements. For example you can now write on your friend’s walls and upload photos stored in your photo library.

As a pretty regular user of Twitter I couldn’t resist getting Twitterific by IconFactory. I’ve been pretty impressed with it so far, and it’s got a rating of around five stars on the App store. Twitterrific will show you your timeline of tweets with the ability to reply to and direct message people you are following. You can also access your friend’s profiles and star tweets. All round it does everything you need.

There is a premium version of Twitterrific, which gives you the option of a lighter theme and no adverts but to be honest I haven’t found the ads to be too obtrusive so haven’t opted to buy the premium version.

I’m loving the Last.fm applications because there’s nothing quite like free music and Last.fm’s recommendations are always really good! If you’re not already using Last.fm sign up today! You can share your musical tastes with other people and find loads of guys and gals who listen to the same embarrassing rubbish as you! I’ll post a load more about Last.fm in the next couple of days because there’s more to it than meets the eye…

So now you’re posting pictures to your friends on Facebook, tweeting as you sit on the train to work and listening to internet radio tailor made to your musical tastes. But wait, you decided to pop in to a trendy little independent clothes store to pick up a pair of jeans for the weekend and you hear this great track playing in the background. Those rolling lo-fi beats would make a perfect addition to your pre-partying warmup mix but you have no idea who the song is by or what it’s called. There are after all no words!

Out comes the iPhone with Shazam. Shazam is amazing and it’s currently free! This app listens to what’s going on around your iPhone, picks up a song in the background and tells you what it is. And it works!

I’ve only had it fail on me once and that was when the music was really quiet and there were loads of people talking. To be fair, I could barely hear the music myself! It did, however, manage to recognise an Al Green track in Pulp Fiction while in the midst of onscreen dialog. Overall, a must have app!

Tap Tap Revenge is the only real game I have on my iPhone. It’s kind of like a guitar hero dance mat kind of game where you tap the screen or shake your iPhone in time with music. It’s quite a lot of fun especially when playing two player head to head mode! And there are additional tracks available for download so the shelf life of the game can be extended.

I’ve played Super Monkey Ball on the iPhone and was quite impressed with the logistics of it all. It makes great use of the accelerometer and feels well made. The only reason I haven’t bought it is I can’t justify spending £5.99 just to waste all of my battery life in an hour.

The next app I want to try out has to be the VNC viewer. I don’t think it would be that practical in every day situations. For example, I could never use it for controlling a server over a VPN connection, it would just be far too slow; and I’d rather use SSH anyway. But it does seem like a nice idea when I’m floating around the house controlling iTunes remotely. And it would be quite good at showing off what the iPhone can do.

When I get my 3G I’ll no doubt be able to rethink some of my app choices as the size of my pipe will increase massively. Until then, I’ll just stick with my edgy little pipe and use bandwidth hungry apps over WiFi only.

When working with colour in Terminal we have to use a rather cryptic series of escape characters with colour codes to present the user with some beautiful letters.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

#!/usr/bin/env ruby -wdK
 
esc = "\033["
fgs = (30..37).to_a
bgs = (40..47).to_a
box_width = 4 + 4 * bgs.length
 
puts "", "Ruby colour box".center(box_width), ""
 
print " "*4
 
bgs.each_with_index do |bg, i|
  fg = fgs[i]
  print " #{bg} "
end
 
print "\n"
 
fgs.each do |fg|
  print " #{fg} "
  bgs.each do |bg|
    print "#{esc}#{fg};#{bg}m #{fg} #{esc}0m"
  end
  print "\n"
end
 
puts "", "http://www.theonlyjames.com".center(box_width), ""

The Ruby script illustrates the structure pretty well. You have an escape character (in our script it’s defined as “\033″. For convenience I’ve bundled the “[” on the end of my variable) followed by the foreground colour code, a semicolon, the background colour code and finally a closing “m.”

The escape character and the following options are cumulatively called an escape sequence.

After our string we restore the default colours using another escape character followed by “0m.”

There are a number of further codes we can use after our escape sequence to control the behaviour of our shell although most are fairly specific to the environment in which you run them.

Worth noting, something I played with after starting to use colour. I drafted a quick colour check step that could be added to any script before we start printing out illegible escape sequences that might not be interpreted. For example, when testing quickly in TextMate…

1
2
3
4
5
6
7
8
9
10
11
12
13

#!/usr/bin/env ruby -wdK
 
term = %x(echo $TERM).strip!
 
print "\n"
 
if term.downcase == "xterm-color"
  puts "Colour supported in \033[32;40m#{term}\033[0m!"
else
  puts "Unknown environment (#{term})."
end
 
print "\n"

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

#!/usr/bin/env ruby
 
# Name of group to apply
new_group = "teachers"
 
home_folders = Array.new
sub_folders = %w(Pictures Documents Library Music Movies Sites Desktop)
d = Dir.new(".")
 
def do_cmd(*cmd)
  if $DEBUG
    cmd.each {|c| puts c}
    puts
  else
    cmd.each {|c| `#{c}`}
  end
end
 
d.each do |item|
  if File.directory?(item)
    home_folders << item
  end
end
 
home_folders -= %w(. ..)
 
puts "Home folders found: ", home_folders.join(", "), ""
 
home_folders.each do |item|  
  do_cmd("sudo chown #{item}:#{new_group} #{item}", "sudo chmod 770 #{item}")
 
  sub_folders.each do |sub|
    do_cmd("sudo chown -R #{item}:#{new_group} #{item}/#{sub}", "sudo chmod -R 770 #{item}/#{sub}")
  end
end
 
if $DEBUG
  puts "You are working in debug mode. No changes have been made."
end

Any home folder will be owned by its respective user. The name of the folder will of course be the shortname, which made this a whole lot easier. Then it was just a case of picking a group and applying the changes to files and folders we knew were safe to change.