The installation went easily for both ruby gems and ruby, but I decided to take a ’short cut’ and copy all my gems from /Library/Ruby/Gems into my /usr/local/lib directory which started raising all kinds of errors — this one, in particular was obnoxious.

[BUG] cross-thread violation on rb_gc()

Luckily, all that means is that I copied over gems which were compiled against the standard OSX ruby version and not the new one. This was a little script I wrote which will show you which gems need to be re-compiled. Just cd over to your /usr/local/lib/ruby/gems/1.8/gems directory and run:

gems $>ls -1 **/**/*.bundle|ruby -pe '$_.gsub! /\-.*/, ""'|uniq

to get a list, or pipe that into $> sudo gem install and that should clear up those gc issues.

The default YAML output

This is what you will get if you simply throw the YAML.load result to an array or something. Nice, but I’m not a big fan of “Keys Like This” so…

irb(main):011:0> pp YAML.load(`svn info`)
{"Node Kind"=>"directory",
 "Last Changed Author"=>"rob",
 "URL"=>"http://example.com/share/lib",
 "Schedule"=>"normal",
 "Last Changed Rev"=>441,
 "Repository UUID"=>"0afc494f-e74d-0410-99f2-b94b27995134",
 "Repository Root"=>"http://example.com",
 "Last Changed Date"=>"2008-07-15 15:55:54 -0400 (Tue, 15 Jul 2008)",
 "Revision"=>446,
 "Path"=>"."}
=> nil

A nice class wrapper for the SVN Info

Heres what I ended up with. Nothing fancy, just a simple wrapper around the subversion info dump.

require 'yaml'
class SVNInfo
  def initialize
      YAML.load(`svn info`).each do |k,v|
          key = k.gsub(/\s/, '_').downcase
          instance_variable_set "@#{key}", v
          instance_eval %{ def #{key}; @#{key}; end }
      end
  end
end

Here is the end result:

irb(main):013:0> svn_info = SVNInfo.new
# ... snip ...
irb(main):014:0> pp svn_info
<SVNInfo:0x57f378
 @last_changed_author="rob",
 @last_changed_date="2008-07-15 15:55:54 -0400 (Tue, 15 Jul 2008)",
 @last_changed_rev=441,
 @node_kind="directory",
 @path=".",
 @repository_root="http://example.com",
 @repository_uuid="0afc494f-e74d-0410-99f2-b94b27995134",
 @revision=446,
 @schedule="normal",
 @url="http://example.com/share/lib">
=> nil
irb(main):015:0>

Messing Around

irb(main):018:0> pp svn_info.methods - Object.methods
["last_changed_author",
 "revision",
 "repository_root",
 "last_changed_rev",
 "path",
 "url",
 "node_kind",
 "last_changed_date",
 "repository_uuid",
 "schedule"]
=> nil
irb(main):019:0> svn_info.revision
=> 446
irb(main):020:0> svn_info.last_changed_author
=> "rob"
irb(main):021:0>

Hope that comes in handy

If you’re curious to see which commands you use the most just run:

$>history | awk '{a[$2]++}END{for(i in a){print a[i] " " i}}' | sort -rn | head

# local server
133 l
70 cd
47 svn
38 ..
29 e
26 mv
20 ll
10 rm
10 cp
8 ror

# remote server
123 l
74 cd
35 svn
31 touch
26 ..
23 exit
23 cat
20 ll
18 rm
16 rake

A few notes

here are those goofy aliases that you see:

alias ll='ls -lah'
alias la='ls -a'
alias l='ls -lh'
alias ..='cd ..'
alias ~='cd ~'
alias e='mate '

alias ss='./script/server'
alias console='./script/console'
alias migration='./script/migration'

alias ror='/var/rubydev'
alias www='/var/www'

And yes, this needs it own post simply for the fact that this is an insane prompt. Heres what it is/can do:

1. Automatically shortens itself if the path is longer than 18 charachters
   This is what I’ve always wanted since I usually name my folders after the website URL — and I HATE 2 line prompts.
2. Changes to red if the last command didn’t have a return code of 0
3. Includes your bash history # for easy history command repeating
4. Is pretty neat and trim

Why is this rad?

Prompt Madness!

I’ll let this image explain what it does and how it looks…

I’m tossing up a text file here -> best_bash_prompt_ever.txt — since the escaping will probably be messed up.

export PROMPT_COMMAND='PS1="\[\033[0;33m\][\!]\`if [[ \$? = "0" ]]; then echo "\\[\\033[32m\\]"; else echo "\\[\\033[31m\\]"; fi\`[\u.\h: \`if [[ `pwd|wc -c|tr -d " "` > 18 ]]; then echo "\\W"; else echo "\\w"; fi\`]\$\[\033[0m\] "; echo -ne "\033]0;`hostname -s`:`pwd`\007"'

If you think you’re interested — this post is about the following:

1. How to get your current user/directory in the title of terminal (iTerm.app for this post — same applies for Terminal.app though)
2. How to make the backspace/delete keys work properly in nano/pico on local and remote servers
3. Update your current user/path in the title when you SSH into a remote machine
4. And how to setup password-less SSH, just because its on topic

.bashrc modifications

Making Terminal.app’s title update to the current user & path for both local and remote servers while letting the backspace and delete key work properly — all in one solution!

Add the following lines to both your local & remote .bashrc

# this is what updates your Terminal's title -- you can of course modify this
# but im not explaining how -- google it
export PROMPT_COMMAND='echo -ne "\\033]0;${USER}:${PWD}\\007"'

# this will make OSXs backspace & delete keys work properly
# in nano & pico
export TERM=xterm
stty erase ^H

# fancy prompt you say?
# again -- google for more info on this
export PS1="\\[\\033[0;34m\\][\\!]\[\\033[0;32m\\][\\u.\\h: \\w]\[\\033[0;32m\\]\\$\\[\\033[0m\\] "

#gives you something like:
# [0][rob.hades: ~]$ echo 'hello world!'

Now lets bind our servers for SSH so we don’t need passwords

Passwords are annoying — especially when you are using scp and rsync a lot.

Here is a very simple way to drop those passwords and have SSH do all the work.

In your terminal app (Locally)

Make your .ssh directory
$>mkdir -p ~/.ssh && cd ~/.ssh

Generate yourself some keys
~/.ssh $>ssh-keygen -t dsa -f ~/.ssh/id_dsa

Copy your public key to your remote server (Make sure the remote server has a ~/.ssh directory as well!)
~/.ssh $>scp id_dsa.pub {user}@{remoteserver}:.ssh

In the remote computer

Add that public key to your authorized_keys file for SSH
$>cd .ssh && cat id_dsa.pub >> authorized_keys2

Cleanup those files and make sure the authorized_keys file is set to the right perms.
~/.ssh $>chmod 640 authorized_keys2 && rm id_dsa.pub

That should make your life a bit more convenient

Using pbcopy and pbpaste can save your life sometimes and are probably one programs I miss the most when using Linux. Anyway, heres just a simple “1 liner” for prepending.

cat FILE | pbcopy && echo "Prepend Text" > FILE && pbpaste >> FILE

Pretty simple, but it can be useful at times — like when you have a script that would be better off with a shebang line than being run by hand and you’re too lazy to open an editor.

Problem

Use Linux’s “find” command to hunt recursively through the current directory tree and find all matching files. While this is simple enough, I’m an idiot and tend to delete important files when drinking working too much — which we all know sucks. It would be nice to have it take all found files, tar them up and move it to the trash.

Solution

The evil, but oh-so-useful xargs command is your saving grace. While I hate using xargs its damn useful. If you take the output of find and pipe it into tar then pipe THAT into rm you get the job done.

Example

This is just a simple hack I’ve put together which seems to work and is relatively simple. (done as a bash function)

frm(){
	find . -name "$1" | xargs tar rvf "$1.tar" | xargs rm -rfv && mv $1.tar ~/.trash
}

I just needed this because I had to wipe all ./.svn directories so i just typed frm .svn and it created a .svn.tar and moved that to ~/.trash. Good stuff

Gnome’s default keybinding app sucks (maybe Linus was right and Gnome treats its users like idiots? sucks for me I guess). Here is a way to bind anything you’d like to any key you’d like, globally. I found this useful because I naturally use WIN-E for an explorer or WIN-R for a run dialog and I couldn’t get that working the right way through Gnome’s default shortcut app (/System/Preferences/Keyboard Shortcuts).

Anyway, doing this is really simple. Open up a terminal and type gconf-editor. Browse the /app/Metacity/global_keybindings and keybinding_commands. Edit the corresponding global_keybinding command_# to the keybinding_command command_# and you’re golden.

Click for larger view

Click for larger view

Step 1:
In gnome goto System -> Administration -> Network and click on the ‘Hosts’ tab

Step 2:
For your normal localhost (127.0.0.1) double-click or hit ‘properties’ to edit it. Add your list of subdomains under localhost.

*Ignore the text in this pic, I overlooked that you had to include _all_ aliases in the same box. Whoops

Step 3:
The fun step
Open a terminal and type sudo gedit /etc/apache2/sites-available/default — assuming you’re using the “default” httpd.conf. Chances are that you will be, and if you aren’t you should know what file I’m talking about anyway. (If you don’t, just load one of the enabled apache configuration files)

Change these lines:

NameVirtualHost *
<VirtualHost *>

to

NameVirtualHost 127.0.0.1
<VirtualHost 127.0.0.1>

Step 4:
Add the following (for each subdomain you made) — changing the ServerName to whatever you aliased in the network host dialogue and DocumentRoot to wherever you want the subdomain to point.

<VirtualHost 127.0.0.1>
	ServerName mysql.localhost
	DocumentRoot /var/www/phpmyadmin
</VirtualHost>

Step 5:
In a terminal type: sudo apache2ctl restart
When that finishes, browse to http://mysql.localhost and you should see phpMyAdmin (going by my example)

On OSX + Apache 1.33

Step 1:
Open /Applications/Utilities/NetInfo Manager and click on “machines”
I usually just duplicate the localhost settings, or just create a new record below localhost that is something like:

ip_address: 127.0.0.1
name: mysql.localhost
serves: ./local

Click for bigger version

Step 2:
Repeat Steps 3-5 above only editing “/etc/httpd/users/{yourname}.conf” instead.

Click for bigger version

Restart Apache and hopefully all went well. I found this useful because I hate typing unnecessary paths

Theres a few common solutions:

Every Flavour Beans mentions its either a login timing (GDM) issue, or a samba auto-mount issue. The solution there was to either: a) disable all samba shares (I’m not 100% if they meant temporarily or forever — either way, this was a last resort since I hate this kind of stuff) and b) disabling automatic logins through gnome’s GDM.

I didn’t have auto-login enabled, so I figured “what the hell” and set it to auto-login with a 10 second timer. When I re-logged I didn’t get the HAL error, but I also didn’t get my auto mounted drives back. Shit. Next.

I decided to poke around a bit more to see what the bigger picture is. I knew it couldn’t be a big issue since I wasn’t installing/configuring/etc. at the time of the crash — something just got “stuck.” I stumbled upon another site which mentioned running sudo /usr/sbin/hald --daemon=no --verbose=yes to view the output.

After running that I got a suspicious message pointing to the real issue: Error binding udev_event socket: Address already in use. Damn udev. Anyways, I tossed up a sudo /etc/init.d/udev restart and re-booted. Lo and behold – all is well.

(Again, this post is more for my future reference in case I get this error again)