Search My Blog

ruby (4) web (4) ruby on rails (3) security (3) GPG (2) OpenPGP (2) RFC (2) linux (2) rails (2) shell (2) sysadmin (2) Exchange (1) GIT. (1) IMAP (1) RCS (1) SSH (1) SVN (1) bundle (1) cURL (1) command line (1) crack (1) css (1) developer (1) email (1) fail (1) hack (1) http (1) mac (1) network (1) password (1) regular expression (1) script (1) subversion (1) terminal (1) textmate (1) tip (1) vim (1)

Tuesday, October 25, 2011

Ruby how to get my private and public IP address



Ok,
you've just deployed your newest ruby app on a bunch of servers, and you need that this app knows the IP address of the server where it's running.

I've read some bizarre ways ("Get your local IP address" or "Get your local IP address") to do this, such as opening an UDP socket and inferring it from the interface used to route the packet.
With the Socket class you may do this more easily and also get the benefit of having useful Addrinfo objects and you are able to distinguish easily between public and private interfaces.

First of all:
Socket::ip_address_list
returns the Array of Addrinfo objects with all your interfaces (it deals with both IPv4 and IPv6).

You can then filter them using the standard Enumerable methods select() and detect() along with these Addrinfo methods:
Addrinfo#ipv4?
Addrinfo#ipv6?
Addrinfo#ipv4_private?
Addrinfo#ipv4_loopback?
Addrinfo#ipv4_multicast?
and convert them to string in dotted notation with
Addrinfo#ip_address

as in:

def all_my_ipv4_interfaces
socket.ip_address_list.select{|intf| intf.ipv4?}
end

def my_loopback_ipv4
socket.ip_address_list.detect{|intf| intf.ipv4_loopback?}
end

def my_first_private_ipv4
socket.ip_address_list.detect{|intf| intf.ipv4_private?}
end

def my_first_public_ipv4
socket.ip_address_list.detect{|intf| intf.ipv4? and !intf.ipv4_loopback? and !intf.ipv4_multicast? and !intf.ipv4_private?}
end

all_my_ipv4_interfaces
=> [#<Addrinfo: 127.0.0.1>, #<Addrinfo: 10.212.142.73>, #<Addrinfo: 62.127.4.80>]
my_first_private_ipv4
=> #<Addrinfo: 10.212.142.73>
my_first_public_ipv4.ip_address
=> "62.127.4.80"

Monday, October 10, 2011

Real citizens of a virtual world or virtual citizens of a real world?

I like GPG.
When I started wandering around the Web with my modem about fifteen years ago, the impression was that of being a ghost. There were channels and newsgroups, but behind the words you wrote you could be anyone.
Internet was anonymous.
Time has passed, at the beginning we worried about not being anonymous anymore, then we started not wanting to be that anymore. The Internet has become a virtual extension of our social space.
With the first social networks we thought we could finally have the nationality of this virtual world, but once again we were wrong. We have become virtual citizens of a real world.
What we do on social networks has real world consequences because, in fact, the Internet IS the real world.
Every company has its own site, unique and recognizable. It is difficult for a phishing or cybersquatting not being soon discovered.
But, on the Internet we are less real than the Internet itself. Our virtual identities are ephemeral and too easy to counterfeit and violate. Anyone can pretend to be us, by registering on our behalf, robbing a password, or self attributing pictures, videos, comments or even entire blogs.
And increasingly those who do not know us personally take an idea of us with an online search.
Yet there is a standard protocol (RFC 4880 ), a standard as the email and the Internet itself, which guarantees to each of us a Pretty Good Privacy (PGP).

Each sysadmin knows and uses SSH. And being lazy as all the sysadmins has learned that he can store his public key on remote servers for not even having to type a password to connect.
Indeed, this mechanism should provide better security than passwords, but it's not true because nobody cares about the keys and keeps them safe. On the contrary, during hardware or software changes SSH keys are easily regenerated.

A sysadmin has learned that this message:
means to delete the corresponding line from ~ / .ssh / known_hosts
Almost all sysadmins that I know, at the sight of the message, not even go look for the line and delete the entire portfolio of keys.
And even the very rare cases of people who care about preserving and checking the keys, completely ignore their AUTHENTICITY.

GIT is a distributed code versioning system, and for programmers is a revolution. But lacking the central server repository as a guarantee of the revisions, disappears even the last glimmer of authenticity of the code.
So we put it all on GitHub ... but as long as we rely on the SSH only we are just at the same point.

GPG is not more complex than GIT. Those who keep care of their GPG keys why don't use this portfolio of keys for SSH?
GPG is powerful. Allows you to generate subkeys of the primary key (which should be kept on a disconnected storage and used only when necessary), to choose an expire date, and even to revoke them.
And unlike certificates is as reliable, is free and requires no bureaucratic times.
I have read of the possibility to export a GPG subkey and use it as public key for SSH, but publications on the correct procedure are scarce. Since when it comes to safety it's better not to improvise, I decided to ask for help from someone who was an expert rather than do it by myself. On StackOverflow the question was even banned from a security specialist arguing that it "solicited opinions, debates, discussions, surveys, or flaming".
Maybe no one really cares, because we all want to remain virtual citizens of a virtual world.



Read this post in italian

Thursday, October 6, 2011

Add a shortcut to TextMate to lookup a class or method definition in a tooltip

UPDATED Mar 11 Ott 2011 12:52:36 CEST

I'm Vim user. I love the command mode, and I had fun to dance with those keystroke commands.
However changing OS means viewing things in another perspective, and switching from Linux to Mac OS for me meant choosing an editor that has its roots in the operating system where it runs.
In the case of Mac OS it not only means it should have a suitable GUI, but it should fire up instantly and above all it should behave like a Mac App.
TextMate was the obvious choice. Because behaving like a Mac App means that if you pick a folder in the Finder and drop it in the editor icon, it should straight open it like a Project. And it means that under the guise of a bare and simple application hides a world of possibilities and customization. That world in TextMate is called Bundles.

Let's talk about Bundles with a real case. (TextMate|I) was missing a shortcut to quickly lookup the definition of a class or method name, so I decided to write a TextMate command that shows me the definition in a tooltip for the current word under the caret.
The difficulties I had were merely related to the complexity of this shell scripting. Since I was trying to keep it as simple and efficient I could, I chose to use the standard unix tools: grep find and sed (it ended up that maybe a ruby script would have done it easier :))

All you have to do to use it in your TextMate is to open the Bundle editor (⌃ ⌥ ⌘ B), make a new command under "Ruby", assign a "Key Equivalent" like ⌃ ] and paste the following script in the "Command(s):" windows
#!/bin/bash
SEARCHPATH=$(test -x "$TM_PROJECT_DIRECTORY" && echo "$TM_PROJECT_DIRECTORY" || echo $(dirname "$TM_FILEPATH") )
STARTPATTERN="^([[:space:]]|;)*(def|class)([[:space:]]|;)+$TM_CURRENT_WORD([^A-Za-z0-9_].*)*$"
ENDPATTERN="^([[:space:]]|;)*end([[:space:]]|;)*"
FILESMATCH=$( (egrep -ns "$STARTPATTERN" "$TM_FILEPATH" || egrep -nsR "$STARTPATTERN" --exclude="*.svn*" "$SEARCHPATH") |cut -d: -f-2 )
echo $FILESMATCH # Show on tooltip
echo $FILESMATCH |cut -d: -f1 |xargs -0 basename |pbcopy # Copy on clipboard for later use with cmd+T
find "$TM_FILEPATH" "$SEARCHPATH" -type f -name '*.rb' -exec sed -En "/$STARTPATTERN/,/$ENDPATTERN/p" {} \; |sed -E "/$ENDPATTERN/q"
CODE UPDATED Mar 11 Ott 2011 12:52:36 CEST
Now it also copies the first matching filename to the clipboard, so you can open this file by pressing cmd+T (and then go to the exact line with cmd+L)

Make sure the output is "Show as Tool Tip" like this:

If you've done it correctly, placing the caret on or after any method or class name and pressing ⌃ ] (or whatever key assignment you've made) will search in the current file ($TM_FILEPATH) and then recursively in the project directory ($TM_PROJECT_DIRECTORY) or in the file directory (if file is not in a project), and show a tooltip with:
  1. the name of the file (if different from the one you are editing) and the line where the first definition of the method or class is found
  2. the portion of definition from the declaration to the first "end" found

Although not perfect, it will do its best to avoid commented lines and false matches.
Even if this script was done for ruby, you can use it as a base for other programming languages, just change the -name '*.rb' part of the find command in the extension of your sources, and the keywords def - class - end that matches your method, function or class definitions.

In the very unsual case you are running TextMate not on a Mac OS, or if you want to adapt the script to run in a linux shell, just skip the line to "Copy on clipboard for later use with cmd+T" and remember that to use the Extended Regular Expression in SED you have to use the -r switch instead of -E.

For any explanation, contribution or discussion on the script please leave a comment below, and if you find it useful please leave me a positive feedback! :)

Tuesday, October 4, 2011

My CSS stylesheet for Ruby on Rails syntax highligthing

As my very first post, I would like to share with you the CSS stylesheet I use to format Ruby snippets.
If you need a CSS stylesheet ready to be applied to your html code block, or you simply like my ruby formatting color schema, you can take it!

div.rb-code {
 padding:1em 1em 1em 1em;
 border:0.2em groove #000000;
 font-family:Andale Mono,monospace;
 font-size:1.2em;
 color:#CC9900;
 background-color:#3A3A3A;
 overflow:auto;
 z-index:1;
}
div.rb-block-def {}
div.rb-block-exe {}
div.rb-code div.tab { padding-left:1.4em; }
div.rb-code .attribute { color:#FFFF66; }
div.rb-code .argument { color:#FFFF00; font-style:italic; }
div.rb-code .class-parent { color:#00FF00; }
div.rb-code .class { color:#FF00FF; font-weight:bold; }
div.rb-code .comment { color:#FFFFFF; }
div.rb-code .constant { color:#FF9966; }
div.rb-code .keyword { color:#FF8000; text-decoration:underline; }
div.rb-block-def .method { color:#FF00FF; }
div.rb-code .module { color:#A020F0; font-weight:bold; }
div.rb-code .number { color:#00FFFF; font-weight:bold; }
div.rb-code .regexp { color:#FF0000; }
div.rb-code .string { color:#99CCFF; }
div.rb-code .symbol { color:#CCFF66; }
div.rb-code .var { color:#FFFF66; font-weight:bold; }
div.rb-block-exe .command { color:#99CCFF; font-weight:bold; }
div.rb-block-exe .evaluation { color:#F0F0F0; font-size:0.8em; }
div.rb-block-exe .prompt { color:#CC9900; }

To use it, simply place all your code snippet between <DIV class='rb-code'>...</DIV>, and then wrap inside the appropriate <SPAN>...</SPAN> block the portions you need to highligth.

You can distinguish def blocks and executable blocks placing them inside a <DIV class='rb-block-def'> or a <DIV class='rb-block-exe'>

To ident a block, put it inside another <DIV class='tab'>

Look at the following example:

<DIV class='rb-code'>  
<DIV class='rb-block-def'>
<DIV><SPAN class='keyword'>class</SPAN> <SPAN class='class'>Person</SPAN></DIV>
<DIV class='tab'>
<DIV><SPAN class='keyword'>def</SPAN> <SPAN class='method'>speak_now</SPAN>(<SPAN class='argument'>thought</SPAN>)</DIV>
<DIV class='tab'>
<DIV>puts <SPAN class='var'>thought</SPAN></DIV>
</DIV>
<DIV><SPAN class='keyword'>end</SPAN></DIV>
</DIV>
<DIV><SPAN class='keyword'>end</SPAN></DIV>
</DIV>
</DIV>

it will be shown as:

class Person
def speak_now(thought)
puts thought
end
end