We Heart Code

Easily Add PropertyChangeSupport to Beans in Eclipse

Tim — Tue, 29 Apr 2008 18:58:53 +0000

If you plan on using JFace Databinding in SWT you will have to implement Property Change Support on all your Java Beans, (barf). I wasn't able to find a plugin that would do this in one click to a class so I came up with two ways to do it.

Here's a code template you can use based on the cool trick I learned at Stuff That Happens

firePropertyChange("${enclosing_method_arguments}", this.${enclosing_method_arguments}, ${line_selection});

Just select the assignment part of your setter and use this template.

If you're really lazy you can use a regex find and replace (With the File Search Dialog)

Your Find string will be:
(this.+) = (\w+);

And your replace string will be:
firePropertyChange("$2", $1, $1 = $2);

Both of these solutions assume you use an Abstract BaseClass to set up your firePropertyChange stuff, then extend for each bean.

Mine looks like this:

package com.weheartcode;

import java.beans.PropertyChangeListener;
import java.beans.PropertyChangeSupport;

public abstract class BaseEntity {
private transient PropertyChangeSupport propertyChangeSupport = new PropertyChangeSupport(
this);

/**
   * Adds the property change listener.
   *
   * @param listener the listener
   */
public void addPropertyChangeListener(PropertyChangeListener listener) {
propertyChangeSupport.addPropertyChangeListener(listener);
}

/**
   * Adds the property change listener.
   *
   * @param propertyName the property name
   * @param listener the listener
   */
public void addPropertyChangeListener(String propertyName,
PropertyChangeListener listener) {
propertyChangeSupport.addPropertyChangeListener(propertyName, listener);
}

/**
   * Removes the property change listener.
   *
   * @param listener the listener
   */
public void removePropertyChangeListener(PropertyChangeListener listener) {
propertyChangeSupport.removePropertyChangeListener(listener);
}

/**
   * Removes the property change listener.
   *
   * @param propertyName the property name
   * @param listener the listener
   */
public void removePropertyChangeListener(String propertyName,
PropertyChangeListener listener) {
propertyChangeSupport.removePropertyChangeListener(propertyName,
listener);
}

/**
   * Fire property change.
   *
   * @param propertyName the property name
   * @param oldValue the old value
   * @param newValue the new value
   */
protected void firePropertyChange(String propertyName, Object oldValue,
Object newValue) {
propertyChangeSupport.firePropertyChange(propertyName, oldValue,
newValue);
}
}

Bookmark to:

Automatically Add JavaDoc Comments in Eclipse

Tim — Fri, 11 Apr 2008 13:05:44 +0000

Recently while preparing to contribute some code to Eclipse I had to add a copyright comment to the top of every class. Enter JAutodoc it will go through and add Javadoc comments and other automated comments, the update site is:
http://jautodoc.sourceforge.net/update/

Screenshot:

Bookmark to:

Limiting log4j SMTPAppender

Tim — Fri, 15 Feb 2008 14:43:21 +0000

I wanted to set a maximum amount of e-mails that SMTPAppender would send after filling up a mailbox with 30,000 damn e-mails.

Turns out it was pretty damn easy, I set up my appender in a java class, and the SMTP appender has a method called setEvaluator that takes an instance of TriggeringEventEvaluator, so I just created a quick inner class like so.

class LimitingEvaluator implements TriggeringEventEvaluator
{
private int sentEvents = 0;
private int MAX_EVENTS = 20;
@Override
public boolean isTriggeringEvent(LoggingEvent arg0) {
sentEvents++;
if(sentEvents>= MAX_EVENTS)
return false;
return true;
}

}

and when i set up my appender I do this

appender.setEvaluator(new LimitingEvaluator());

Bookmark to:

Mass E-mailing, worth paying for a service

Tim — Thu, 17 Jan 2008 14:20:48 +0000

Most of us have written little mass e-mailing scripts in the past. No big deal, connect to SNMP, loop through our subscriber list, do our thang. But then the requirements come in: "

"Hey, I need to know what person is clicking on that link, and what time they opened it."

"How does it look in a hotmail account?"

"Does it pass every spam filter?"

Sure, we could code a little link pass through tool, encoded with the person's e-mail address we sent it to, we could create test accounts at all the major e-mail services, we can make all sorts of intricate unsubscribe reports. But why would you want to do that? That's Amateur Hour! You've got cool shit to build. The solution? CampaignMonitor.com. They charge $5 base fee for every "Campaign" you send, plus 1 cent per recipient. So to send to 45,000 people that costs you $50. Big woop.

Plus you get all sorts of intricate reports

They allow you to send full HTML with no size limit, you're not limited to their templates. You also get a nice list manager with unsubscribe and subscribe features built right in.

For a $10 fee you can preview your design in like 10 different e-mail clients & services as well as test that your e-mail passes spam filter (hint: don't put the word viagra in your e-mail).

CampaignMonitor also has relationships with major ISP's so you know your e-mail will not be denied and your servers won't be blacklisted. Plus the office/marketing drones love the reports and the flexibilitiy to send an e-mail campaign themselves.

So don't waste time re-inventing the wheel dummy, pony up your whopper money and do it the easy way.

And no -- they didn't pay me anything for this post, though they can if they want!

Bookmark to:

Postgres select replace with regular expression

Tim — Wed, 02 Jan 2008 21:22:50 +0000

Postgres seriously kicks ass, here's a quick way to do a substring based on a regular expression. A Co-worker wanted to get the first word out of a column that contained a sentence. If we were using something lame like oracle we'd have to do gymnastics to do that.

It's a one-liner in postgres.

SELECT substring('this is a sentence' FROM '^(.+?)\\s.+$')

So we're selecting out the group that is a series of characters at the start of a line (column) before a space. Epic win.

Bookmark to:

Upgrade to Gutsy Gibbon Fast

Paul — Fri, 19 Oct 2007 03:29:34 +0000

Everyone is excited that Ubuntu 7.10 is finally out. (Have a look at the new feature tour here.)

But the servers are getting hammered, so most people are waiting on the upgrade (voluntarily or otherwise).

No need to wait! The more people who download from Bittorent, the faster your speeds! Grab it while it's hot!
Download a copy of the alternative i386 CD here via Bittorent.
Mac and 64 bit torrents can be obtained here.

Once you have the alternative CD, burn it, boot from it, select the upgrade option, and away you go.

Want to be environmentally friendly and save a CD? Try this:

Mount the ISO to /cdrom:
sudo mount -t iso9660 ubuntu-7.10-alternate-i386.iso /cdrom -o loop

Using the full path and proper name for the ISO image.
Open a terminal and type:
sudo /cdrom/cdromupgrade

Then the upgrade utility should launch and run from the CD.

Make sure you have the alternative CD and NOT the standard desktop CD.

Good luck and happy upgrading!

Bookmark to:

Reading An Excel File With Ruby

Tim — Fri, 05 Oct 2007 13:31:35 +0000

This tutorial will cover how to read (or parse) an excel file with ruby. I had to write a script to unpivot some data for a co-worker that saved him hours of time, and I got to write a ruby script, so it was a win-win. Here's how you can do the same thing.

Installing Parseexcel

Parseexcel is a ruby port of the perl parseexcel module.

It's installable via a nice gem like so:

gem install parseexcel

That's that, now remember since it's a gem library we have to tell our script to use the gem libs when we run the script from the console using the -rubygems switch.

Using Parseexcel

Parseexcel is a very straight forward library, you can't do too much with it, but it gets the job done.

Getting a Workbook

Spreadsheet::ParseExcel.parse(filenameandpath)

This returns the actual excel file's workbook, from there we need to determine what worksheet we're on.

Getting a Worksheet

worksheet = workbook.worksheet(0)

Will return the first worksheet, you could also use the each method on the workbook to iterate over all the worksheets.

Iterating over rows and columns

The worksheet object has a very nice each method that will allow us to iterate over the rows like so

worksheet.each { |row|
j=0
i=0
if row != nil
row.each { |cell|
if cell != nil
contents = cell.to_s('latin1')
puts "Row: #{j} Cell: #{i} #{contents}"
end
i = i+1
}
j = j +1
end
}

Getting Cell Data

Getting a String: cell.to_s('latin1')
Getting a Float: cell.to_s('latin1')
Getting a Int: cell.to_i
Getting a Date: cell.date

Getting A Specific Cell

cell = row.at(3) #returns cell at column 3

A basic script for dumping an excel file

require 'parseexcel'

#Open the excel file passed in from the commandline
workbook = Spreadsheet::ParseExcel.parse(ARGV[0])

#Get the first worksheet
worksheet = workbook.worksheet(0)

#cycle over every row
worksheet.each { |row|
j=0
i=0
if row != nil
#cycle over each cell in this row if it's not an empty row
row.each { |cell|
if cell != nil
#Get the contents of the cell as a string
contents = cell.to_s('latin1')
puts "Row: #{j} Cell: #{i}> #{contents}"
end
i = i+1
}
end
}

To run the script, remember to use the -rubygems switch so that you can find the parsexcel library.

ruby -rubygems excelparse.rb myfile.xls

Bookmark to:

Easy XML Generation with Ruby and ERB

Tim — Tue, 18 Sep 2007 15:42:25 +0000

Ah, the bracket tax. XML-Situps, whatever you want to call them, if you write any code (especially if you use java) it should be your sworn enemy. The other day I had to create about 40 XML files for some Java Webstart configs. Instead of doing them by hand, I decided to check out the ERB class for easy "templating" that's built into Ruby. It makes doing code generation very easy!

Our Task

Create a custom XML file for each .jar file in a directory.

...more after the jump!

Strategy

With any code generation project we need to identify a few things.

Static template: The code/text that will stay the same.
Dynamic variables: These are the pieces that will change in our template.
Input source: The input for our dynamic variables.

For me, the static template is a JNLP XML file that looks like this:

<?xml version="1.0" encoding="UTF-8"?>

spec="1.0+"
codebase="http://redacted.com/libs/com.redacted.dao_1.3"
href="RcpWebStart.jnlp">
>
>RedactedLibs>
>Redacted>
>>com.redacted.dao_1.3.1.jar>>
>
>
-permissions/>
>
>
version="1.4+"/>
href="com.redacted.dao_1.3.1.jar"/>
>
-desc/>
>

The dynamic bits will be anywhere a jar is referenced as well as the directory for the jar.

The Template

The template defines our static text, and the variables and/or ruby code that will be substituted into the text by ERB.

We use

<%= %>

blocks to substitute in variables from our ruby script.

Here's what my template looks like, I have four variables that I want to substitute in.

basedir: the base directory of where I'm storing my jar.
shortname: the name of the jar sans the version.
jarname: the full name of the jar file.

<?xml version="1.0" encoding="UTF-8"?>

spec="1.0+"
codebase="http://redacted.net/libs/<%=basedir%>/<%=shortname%>_<%=version%>"
href="RcpWebStart.jnlp">

</span><span style="color: #009900;"><%=basedir%></span><span style="color: #009900;">
Redacted
<%=jarname%>

1.4+"/>
<%=jarname%>"/>
>
-desc/>
>

The Code

Let's examine some code. The first thing I do is create the template string, i Use the %q{ shortcut to avoid having to worry about quotes.

Next, I need to define my template, that's what this line does.

jnlp = ERB.new(template, 0, "%<>");

It says, create a new ERB object using the string 'template' as the template and the "%<>" as my template escape operators.

Our next step is to loop through all the jar's in a directory and create a new file for each one, here's the juicy bits to accomplish that.

b = binding
jarname = File.basename(path)
basedir = ARGV[1]
shortname = jarname.gsub(/(.+?)_(\d.+)$/,'\1')
version = jarname.gsub(/(.+?)_(\d\.\d)(.+)$/,'\2')

dirname = shortname + "_" + version
ourfile = jnlp.result(b)

Notice that we create a binding variable called b and then pass that to our ERB object (jnlp), that binding variable tells ERB to use template variables in the same scope as it. It's some ruby black magic that makes our lives easier.
The rest of the script actually creates some directory and writes the created file to a directory based on the name of the jar.

require "erb"
require "find"

template = %q{
"1.0" encoding="UTF-8"?>
spec="1.0+"
codebase="http://redacted.net/libs/<%=basedir%>/<%=shortname%>_<%=version%>"
href="RcpWebStart.jnlp">

<%=basedir%>
Redactedl
<%=jarname%>

"1.4+"/>
"<%=jarname%>"/>

}

jnlp = ERB.new(template, 0, "%<>");

#Search for jar files here
Find.find(ARGV[0]) do |path|

if !FileTest.directory?(path)
b = binding
jarname = File.basename(path)
basedir = ARGV[1]
shortname = jarname.gsub(/(.+?)_(\d.+)$/,'\1')
version = jarname.gsub(/(.+?)_(\d\.\d)(.+)$/,'\2')

dirname = shortname + "_" + version

#pass in the binding variable and create a string from the template
ourfile = jnlp.result(b)

#create a new directory and file, and output the template string to it.
Dir.mkdir(basedir + "/" + dirname)
outfile = File.new(basedir + "/" + dirname + "/RcpWebStart.jnlp","w+");
outfile < end

The fun doesn't stop there

This is an elementary example, but ERB also allows you to embed real ruby code in the template, this allows you to accomplish looping.

So we could pass in an array of person objects and loop through them with a template like this:

>
<%@people.each do |person| %>
><%=person.name%>>
<%end%>
>

Bookmark to:

Killing a process by name in ubuntu

Tim — Mon, 17 Sep 2007 20:52:23 +0000

I have to kill mongrel alot because it's a flea-bitten mutt.

My usual method was to do a ps -aux | grep mongrel , find the pid and kill it!

Annoying and stupid!

Now I just use the nice pkill tool in ubuntu like so.

pkill mongrel

boom! All bad doggies killed! This probably works on other *nix flavors, but who knows if your stupid admin installed it.

Bookmark to:

Capturing STDERR with ruby backticks

Tim — Fri, 14 Sep 2007 14:13:25 +0000

I was writing a ruby script today that ran a linux command for every line in a file. The problem was the backtick operator in ruby only captures stdout, here's a little trick to get stderr as well. (*nix only)

out = `ourlinuxcmd 2>&1`

the 2>&1 bit tells the shell to redirect stderr to stdout

Bookmark to: