Pound is dead simple to setup and configure. Unfortunately, the darwin port for pound is old and does not work. So this guide will help you build pound on your own. Besides, everyone feels cooler after compiling that hot fire. Click on for the steps.

1. Open terminal.
2. Install zlib

mkdir ~/temp
curl -O http://www.zlib.net/zlib-1.2.3.tar.gz
tar xzvf zlib-1.2.3.tar.gz
cd zlib-1.2.3
./configure --prefix=/usr/local --shared
make
sudo make install

3. Install openssl

curl -O http://www.openssl.org/source/openssl-0.9.8b.tar.gz
tar xzvf openssl-0.9.8b.tar.gz
cd openssl-0.9.8b
./config -L/usr/local/lib --openssldir=/usr/local/etc/openssl \\
    zlib no-asm no-krb5 shared
make
sudo make install

4. Install Pound 2.0 or greater

curl -O http://www.apsis.ch/pound/Pound-2.0.tgz
tar xzvf Pound-2.0.tgz
cd Pound-2.0
sed "s/-o bin -g bin //g" < Makefile.in > Makefile.in.new
mv Makefile.in.new Makefile.in  # Hit y to override any restrictions
./configure --with-ssl=/usr/local/etc/openssl/ --prefix=/usr/local
make
sudo make install

5. Generate an ssl certificate.

cd /usr/local/etc

You can put the next line in ~/.bash_profile if you want openssl available everytime you open the terminal.

export PATH="/usr/local/etc/openssl/bin:$PATH"
sudo openssl/bin/openssl req -x509 -newkey rsa:1024 -keyout our_cert.pem \\
    -out our_cert.pem -days 365 -nodes

Fill out the required information to generate the ssl certificate.

6. Create a pound.cfg file.

cd /usr/local/etc # You should already be here

This next piece uses a trick called a heredoc. Instead of copying and pasting the below command into terminal, you could also just copy and paste the text into pound.cfg in /usr/local/etc/pound.cfg

cat > ~/tmp_file <<EOF
ListenHTTP
  Address 0.0.0.0
  Port    80
  Service
    BackEnd
      Address 127.0.0.1
      Port    3000
    End
  End
End

ListenHTTPS
  Address 0.0.0.0
  Port    443
  Cert    "/usr/local/etc/our_cert.pem"
  # pass along https hint
  AddHeader "X-Forwarded-Proto: https"
  HeadRemove "X-Forwarded-Proto"
  Service
    BackEnd
      Address 127.0.0.1
      Port    3000
    End
  End
End
EOF
sudo mv ~/tmp_file ./pound.cfg

7. Make sure that pound works.

pound -v -c

You should see Config file /usr/local/etc/pound.cfg is OK.
If not, make sure that you copied your config file correctly
using ‘cat pound.cfg’ to view what is in the file.

8. Turn off apache (assuming it’s running).

sudo apachectl stop

9. Turn on pound.

sudo pound -v

10. Start mongrel/webrick/lighty.

cd ~/work/my_killer_app
mongrel_rails start # Or ruby script/server (if you're still on lighty/webrick)

11. Marvel at your wonderous creation by pointing your browser to http://localhost/ or https://localhost/

TODO (feel free to do these and post how to do them in the comments):

• Make pound start when the computer loads.
• Use darwin ports to install zlib and openssl, but not pound.

For those of you not in the know, websense is this stuff that goes between your work computer and the internet. It prevents you from doing cool things like looking at Flickr or downloading songs on iTunes (damn billionth song). It works by maintaining a master database of domain names. You try to access a domain that’s disallowed, it says no. You try to access a domain it doesn’t know about, it sends its little bot out to catalog it. If the bot decides the site is bad, it disallows it.

Now that we know how it works lets look at ways around it. First, we could vnc to a remote computer and use that computer to browse the web. Good idea, but I think that little webcrawler is going to try to access the ip address, notice it can’t do anything with it, and block it. Second, we can try using the google and yahoo caches. These are cool options, but I can’t use them to log into this blog. Also, they’re fairly clunky as far as navigation goes. Third, we could design an open source proxy server that could be easily installed on any computer. If you just try to access the server itself, it could return the google page to psych out the Websense spider. If Websense does manage to catch on to the proxy game, you just setup the server on a different ip. Woo hoo, screw you Websense.

Needless to say, I’m a fan of option three. It sounds cool, and I like to build stuff. Of course, if anyone has a working copy of this option in existence, let me know. As much as a like to build, I’d really like to be posting from work immediately.

My proposed solution? For any programming language that you work in, you should have a utility that allows you to quickly identify variables within a certain scope and to see what type of value they might contain. This would have allowed me to select the function I was working on and notice that there was an extra variable ‘i’ that was initialized as global.

I can think of another situation in which this type of tool would be handy. It could be used to add macro writing capabillites (ala lisp) to any desired language. It already has full knowledge of the constructs of the language and the functions and variables that exist. And so long as the macro could be compiled to the desired language, macro functionality would be available. Maybe you’ll see this sort of tool available here in the future.