Ruby 1.8, resolve symlink recursively.

Ruby 1.9 has added a bunch of niceties in the file class that allow you to find the absolute real path of a file (IE, no symlinks). Here’s a quick snippet that you can use in 1.8 to do the same thing. I’m specifically using this in a utility that will likely be symlimked into /sbin, but i’ve still got to load libraries and configs.

You can replace “$LOAD_PATH” with “$:”, but I find the long form more readable.

file_path = __FILE__
while File.symlink?(file_path) do
  file_path = File.dirname(file_path) +'/'+ File.readlink(file_path)
end
#this is your real path. For example only.
puts "resolved path: #{file_path}"

# add your lib directory to the load path.
# Two directories up, as i'm keeping my primary .rb file in project_dir/bin
# search our libraries first to give priority above system installed libs.
$LOAD_PATH.unshift File.expand_path("../../lib/",file_path)
# require something from your lib directory
require 'mylib/lib.rb'

Snow Leopard & X11 Forwarding

I’ve got a really old Solaris app that requires Xforwarding.

Due to some security policies on the system, ssh -X wasn’t working out for me. I decided to manually set the display on the remote machine to point at a local X server.

I ended up having to make the following changes to the preferences to allow remote display forwarding (not through ssh)

It’s a good idea to toggle the settings back when you’re done.

“The following URL is not valid” when Publishing InfoPath Forms

This issue and various solutions, seems to be pretty rampant on the internet – i’ve just run into it myself

The solution for me was quite simple. The URL you are publishing to must be present in the Alternate Access Mapping (AAM) list of your SharePoint site. I was trying to publish to the IP rather than the server name, which was not in said list.

The URL you are publishing to must be present in the Alternate Access Mapping (AAM) list of your SharePoint site.

I was attempting this strange behaviour due to some temporary naming issues that i’m having in my lab network. While most SharePoint folk will not come across this error, I figured it might be useful to throw in my two cents as other people may tossing together temporary sites with little regard to the proper configuration of AAMs.

One of the comments on this blog post clued me in. The person mentioned missing host headers, at which point I realized that IIS probably doesn’t send them if you’re not requesting a “valid” SharePoint URL, as SharePoint itself probably adds the MicrosoftSharePointTeamServices header which Infopath uses to determine if the publishing target is valid.

VMware Server + Infrastructure Manager

I have a new ESXi box to play with and decided to try to connect to a VMware server machine using the infrastructure client. Lo and behold, it worked!

All you can really do is create VMs and view the consoles, but it works a heck of a lot better than the web interface.

If you’ve got it running on a non-standard port, you can connect to the SSL enabled port via

<ip of machine>:<vmware ssl port>

SSL, Apache and Virtualhosting

I’ve got a few sites sitting on this poor web server, which all need SSL for things like webdav and administrative access. I don’t like the SSL errors that come up on the sites, as they all share an IP. To remedy this, I’ve created my own CA and have signed certificates that are good for multiple domains. I’ve distributed my root certificate to everyone who wants to trust me for the purpose of SSL on my sites, or any other certificates that I happen to sign.
Read more →

Growling your shell in OSX

My buddy at overtherailing.com posted a really cool snippet that he picked up on the internet. I am in turn posting it here:

function n {
  "$@"
  status=$?

  if [ "$status" == "0" ] ; then
    result="completed"
  else
    result="FAILED ($status)"
  fi

  growlnotify -m "Script $result" -s "$@"
}

Non-mac users, eat your heart out.

For bonus points, you can plug this sucker into prowl. I plan to make a standalone prowl equivalent for use on other *nix platforms with cURL support.