Holger Kohnen's Thoughts

("php" == 0) is true

2010-02-24T17:14:00.004+01:00

Believe it or not but this is a feature of the language.

http://www.php.net/manual/en/types.comparisons.php
(look for the second table)

PHP automatically casts strings to number, if the boolean expression contains a number or both of the strings are containing only numbers.

So the expression ("01" == "1") is also true,
but ("0" == "php") is false.

You could use strict equal(===) to avoid this behaviour.
Which is a good idea, if you compare different types
or strings with numeric content.

You can't compare apples with oranges, because they are allways different.
But with php you can do it and even better they are equal...;-)

Testing Websites in older Internet Explorer's

2007-08-27T14:38:00.000+02:00

If you need to test your Websites against older Versions of Microsoft's Internet Explorer this could be the Link you need:
http://tredosoft.com/Multiple_IE
Seems not to work on Vista. So, you might want to install an virtual XP machine.

SQLFairy - The SQL Translator

2007-08-09T11:15:00.000+02:00

I needed to convert a database schema from MySQL to SQLite and found a great tool for doing such schema conversion jobs.

# install sqlt on debian sytems with:
sudo apt-get install sqlfairy
# convert from one schema to another schema with:
sqlt -f DBI --dsn dbi:mysql:dbname --db-user root -t SQLite > sqlite.sql

The above command connects to mysql and reads the mysql schema from the connection and writes the sqlite schema to sqlite.sql. You can easily change this command for other databases. If you want to know if sqlt supports your conversion you can run:

sqlt -l
Parsers:
       Access
       DB2
       DB2-Grammar
       DBI
       DBI-DB2
       DBI-MySQL
       DBI-PostgreSQL
       DBI-SQLite
       DBI-Sybase
       Excel
       MySQL
       Oracle
       PostgreSQL
       SQLite
       Storable
       Sybase
       XML
       XML-SQLFairy
       YAML
       xSV

Producers:
       ClassDBI
       Diagram
       Dumper
       GraphViz
       HTML
       MySQL
       Oracle
       POD
       PostgreSQL
       SQLServer
       SQLite
       Storable
       Sybase
       TT-Base
       TT-Table
       TTSchema
       XML
       XML-SQLFairy
       YAML

The Parsers are needed to read the schema and the Producers are needed to write the schema.
Links
http://sqlfairy.sourceforge.net/

Capistrano multistage

2007-07-29T09:06:00.000+02:00

Capistrano our loveliest tool has got a great plugin which allows us to easily stage releases before they going into production. So you can first deploy to testing and wait for the feedback of the QA/Project Owners and than "push the button" to deploy to production. If you need you could use more stages than testing and production.
http://weblog.jamisbuck.org/2007/7/23/capistrano-multistage

Capistrano 2.0 is real

2007-07-23T07:54:00.000+02:00

Great news. Your favourite tool for automating tasks on multiple machines has released his second version. It is really an awesome tool. It makes you forget physical boundries between systems. Deploying is no pain anymore since capistrano does anything for us.

Software Project Management: Retrospectiva

2007-07-23T07:43:00.000+02:00

I must check out this application, it looks really reasonable.
Introduction

Update
Had some troubles installing retrospectiva on my ubuntu box. The solution was to do a:

rake rails:freeze:edge TAG=rel_1-2-3

...in the retrospectiva app directory.
The source of this problem was that i have installed the rails version:
rails-1.2.3.7116 which seems not work with current stable retrospectiva version, this incompatibility is sadly not detected by retrospectivas environment.rb because the check there checks only for the major and the minor version which was corectly in my case. Otherwise everything noted in retrospectivas Quick Install Guide worked like a charm.

Update 2
This tool is awesome. Love the changeset feature, the easy setup, ticket filter and more. Can't wait to hear what non-programmers say to this easy interface.

Developing Adobe Flash Actionscript on Linux

2007-07-08T12:37:00.000+02:00

There are some possebilities out there. But they matched not my needs. What i simply wanted ist a way to develop Actionscript 1. Yes it's true, i want to develop Actionscript 1 ;-). If it is necessary i would develop AS3 or maybe AS2 but for most my needs AS1 is the perfect choice, it's compatible with JavaScript, dynamic and has it's focus on productivity and not on overestimated aspects like type safety.

Flash 8 is my current Flash-IDE. I have tried to run it with wine but with limited success, it was very slow and buggy. So i tried to run Flash 8 from a virtual Windows XP Machine, it's not as fast as with an native Window XP but it is really fast enough to me. It seems that it's zero point three times slower.

So far so good. As an advanced actionscript coder i don't wanted to code my AS1 inside the Flash 8 IDE, because this slows my hacking workflow about factor 3 or 4 down. This is because of the integrated Editor, which is definitly not worth to work with for more than 10 Minutes. My editor of choice is SciTE or SciTE-st. But i don't wanted to code on the virtual Windows XP. I wanted to use my nicely configured Linux version SciTE-st to code my AS1. But how would i test my movies in the virtual machine, without copying Actionscript files and switching to it.

The solution is simple. I installed a ssh server on Windows XP. Installed Robin Debreuil's Flush. And here we go. I simply written a task which copies the actionscript files with scp to the virtual machine and than executes flush on the virtual machine. Flush tests the movie inside the Flash IDE. Great. The last thing to do was to configure the Editor to run this task if i hit the F5 key. So the workflow is really speeded up. No window switching any more, just edit the Actionscript and hit F5 to see movie and output in the virtual machine.

Links
Flush for Flash 8
http://www.flashfanatiker.de/blog/archives/000051.html
Original Flush Homepage
http://www.debreuil.com/flush/
The Company behind Flash
http://www.adobe.com/
Free ssh server for Window XP
http://www.freesshd.com/
Tool for automating tasks like make or ant, but with an uncircumstantial syntax
http://rake.rubyforge.org/
My favourite Editor
http://scintilla.org/SciTE.html
My favourite Version of my favourite Editor
http://caladbolg.net/scite_st.php

Tips
Configuring freeSSHd has taken some time, this is what i have done:

Check the following settings in the FreeSSHDService.ini file, it contains all configuration you can also access through the ui
SSHNewConsole=0 - i needed to deactivate because i had some problems executing commands via ssh, this was on freeSSHd-Version 1.2.0
SSHPasswordAuth=1 - disables password authentication
SSHPublickeyAuth=2 - set public key authentication to required
SSHPublickeyPath=C:\SOMEPATH\ssh_public_keys - set the directory in which freeSSHd searches for files with public keys. Each file has to be named after it's user.
You need to create an user within FreeSSHd, thats all

Subversion Tips

2007-06-20T10:45:00.000+02:00

Subcommander

The goal of the Subcommander project is to build an easy to use, cross platform (Win32, Unix, MacOSX) subversion GUI client (subcommander) including a visual diff and merge tool (submerge).

Links
http://www.yolinux.com/TUTORIALS/Subversion.html
Usefull tips and links, especiallly for linux. You will find also a list of graphical merge tools.

Skalierbarkeit ganz einfach

2007-06-09T11:11:00.000+02:00

Der Vortrag von Jonathan Weiss auf dem Linuxtag 2007 zum Thema Amazon-S3 und Amazon-EC2 kann auf seinem Blog als mp3 runtergeladen werden. Die Folien stehen dort ebenfalls zum download bereit. Super interessanter Vortrag.

Wow, probably one of the nicest consumer interfaces i have seen

2007-06-07T11:45:00.000+02:00

It's a two finger touchscreen interface:
http://www.microsoft.com/surface/

kde hackers little friend - dcop

2007-06-01T15:36:00.000+02:00

Do you want to script kde applications? Dcop could be the answere. There is an command line interface which is surprisingly called dcop. So for a quick start: Open a shell type: "dcop" and hit enter. You will see applications that have a dcop interface. If you are using klipper, which is normally the case, you can control the clipboard through it's dcop interface. Let's see howto do this. Type: "dcop klipper" and hit enter. You will see the supported interfaces. Type: "dcop klipper klipper" to check out the klipper-Interface. You will see the following interface:

QCStringList interfaces()
QCStringList functions()
QString getClipboardContents()
void setClipboardContents(QString s)
void clearClipboardContents()
void clearClipboardHistory()
QStringList getClipboardHistoryMenu()
QString getClipboardHistoryItem(int i)
int newInstance()
void quitProcess()

Now you can simply call a method like this:

dcop klipper klipper setClipboardContents "Hello from dcop"

This will put "Hello from dcop" to your clipboard.
Note that you can use tab-completition, so if you type: "dcop klipper klipper setC" and hit TAB it will complete to: "dcop klipper klipper setClipboardContent".

Is extreme-programming extreme?

2007-05-31T07:38:00.000+02:00

Here is a nice song from the extreme programming japan user group(xpjug) which answeres the question:

Aha nice - Firebug Console API

2007-05-26T17:20:00.000+02:00

The nice Firebug-Console-API api allows us to control the firebug output console and more direct from scripts.

Read more:
Debugging with Firebug an article by Andrew Dupont.

Deploying with capistrano - The copy method

2007-05-24T12:38:00.000+02:00

Capistrano is a great tool for automating tasks on remote machines. In this article i will show you howto update a static webpage where the server has no svn installed. The only things you need are: a local ruby, rubygems and capistrano installation and ssh access to the remote webserver where the files should deployed to.
You need capistrano 2.0 or 2.0 preview, install the preview with:

gem install net-ssh
gem install net-sftp
gem install highline
gem install -s http://gems.rubyonrails.com capistrano

Similar to make's Makefile's capistrano uses Capfile's. To create a Capfile in the current directory execute:

capify .

Now you can edit the Capfile to fit your deployment needs:

require 'capistrano/version'
# Load the deployment recipe
load 'deploy'
# Give a name for your application
set :application, :website
# Define a server role
role :server, 'yourhomepage.com'
# Define where to store the deployed files
set :deploy_to, '/hp/aa/ab/ne'
# Define where to create the symlink to the current release, normally
# this would be your htdocs or www directory
set :current_path, '/hp/aa/ab/ne/www'
# Define that we want to deploy via copy, this is usefull if your server
# has no subversion client installed
set :deploy_via, 'copy'
# Set the deploy strategy to export, otherwise the files are checked out
# from your repository
set :copy_strategy, :export
# Set your ssh user name and password
# (If you omit the password and did not have setup public key authentication,
# capistrano will prompt you for the password)
set :user, 'your-ssh-username'
set :password, your-ssh-password'
# Set the project repository. All files inside the specified directory
# will deployed.
set :repository, 'file:///repositories/website/public'

# Override tasks not needed for deployment of static files
namespace :deploy do
task :finalize_update do
logger.info 'do nothing - overridden finalize_update'
end
task :restart do
logger.info 'do nothing - overridden finalize_update'
end
end

Thats it, now you can setup your host for deployment with(First you need to go into the directory where your Capfile is stored):

cap deploy:setup

And deploy with:

cap deploy

You will find many default setting in: lib/capistrano/recipes/deploy.rb.

Ressourcen schonende Backuplösung

2007-05-21T12:18:00.000+02:00

Die "Heinlein - Professional Linux Support GmbH" stellt eine Backuplösung vor, welche die Möglichkeiten von rsync, hardlinks und ssh voll ausschöpft. Ich verwende diese Lösung nun seit einigen Wochen und bin sehr zufrieden damit. Die Backups laufen schnell und belasten nicht das Netzwerk.
Hier ein Auszug aus dem Artikel:

Stabiler täglicher Betrieb ohne Wartungsarbeiten
Komprimierter Transfer der Daten im Netzwerk
Bandbreitenlimitierung des Backup-Prozesses
Sicherer Transfer der Daten – ssh-verschlüsselt
Snapshot des gebackupten Systems, Rekonstruktion eines bestimmten Zustandes vor n-Tagen möglich
Bequemes Backup auf einfachen Festplatten, ohne teuren Tape-Roboter oder teure Bandlaufwerke. Große Datenmengen möglich.
Schnelles Zurückspielen der Daten in Festplattengeschwindigkeit möglich (schneller als Bandlaufwerke)

Testing with Rails

2007-05-19T11:09:00.000+02:00

Jay Fields has labeled all his great posts about rails testing with: "RailsConf2007". Now we can find all relevant posts more easily. Thanks.
Source: http://blog.jayfields.com/2007/05/railsconf07-testing-presentation.html

The new Art of Flamewar

2007-05-16T11:23:00.000+02:00

Please take the following not to serious. Flamewar can pull your energy off. I think the best thing we can do is respect other people independent from the technologies they are using. Sometimes it is hard to talk to people when they want to argue you into something. Encounter them with acceptance for there opinion. Do not try argue them into something. Instead explain your opinion only if they really want to know your opinion. Just my two cent.
Rails vs. Java

Rails vs. PHP

Source: http://www.railsenvy.com/

Backup one file

2007-05-12T13:15:00.000+02:00

I found my self often doing something like that:

sudo cp /boot/grub/menu.lst /boot/grub/menu.lst_bu_2007-05-12_13-20

With the following script all we need to do is:

sudo backup-one-file /boot/grub/menu.lst

And the result is a backup created in /boot/grub/menu.lst_bu_2007-05-12_13-21-30

#!/usr/bin/env ruby

require 'fileutils'

if ARGV.empty?
        puts "Usage: backup-one-file <file-name>"
        exit
end

source = ARGV[0]

unless File.exist?(source)
        puts "Error: File #{source} does not exist!"
        exit
end

target = source + "_bu_" + Time.now.strftime("%Y-%m-%d_%H-%M-%S")

begin
        FileUtils.cp(source, target)
rescue Exception => e
        puts "Could not copy to #{target}"
        puts e.message
        puts "Maybe you need to run as root!"
else
        puts "Backup created in: #{target}"
end

ri and ri_proxy

2007-05-09T11:45:00.000+02:00

Yesterday i have written a little caching proxy script for ri. The script forwards any invocation to ri. The results are cached in a sqlite3 database. So that any subsequent calls are much faster, not fast as i wished but definitely faster.

#!/usr/bin/env ruby

require 'rubygems'

ROOT = File.dirname(__FILE__)
BASE_NAME = File.basename(__FILE__)

DB_FILE = File.join(ROOT, BASE_NAME + '.sqlite3')
DB_CONFIG = {:adapter => 'sqlite3', :timeout => 5000, :database => DB_FILE}

LOG_FILE = File.join(ROOT, BASE_NAME + '.log')

# speed up active record loading a bit
require 'active_support'
RAILS_CONNECTION_ADAPTERS = ['sqlite']

require 'active_record'
ActiveRecord::Base.logger = nil # Logger.new(LOG_FILE)
ActiveRecord::Base.establish_connection(DB_CONFIG)

class RiMethodEntry < ActiveRecord::Base
end

force = false
if !RiMethodEntry.table_exists? or force
  ActiveRecord::Schema.define(:version => 1) do
    create_table :ri_method_entries, :force => true do |t|
      t.column :name, :string, :null => false
      t.column :text, :text, :null => false
    end
    add_index :ri_method_entries, :name
  end
end

ARGV.length > 0 or raise Exception.new("No arguments found!")
name = ARGV[0]
entry = RiMethodEntry.find_by_name(name)
if entry
  puts entry.text
  exit
end

args = ARGV.collect { |a| %{"#{a}"} }
text = `ri #{args.join(" ")}`
found = text.starts_with?('-')

puts text
exit unless found

RiMethodEntry.create(:name => name, :text => text)

Install ruby on debian sytems the apt way

2007-05-03T13:00:00.000+02:00

If you want to install ruby through apt you normally get only the interpreter. If you want to install ruby with all it's tools, you need to take a look into the package description to find out what is missing.

apt-cache show ruby1.8

The output should be similar to the following:

Package: ruby1.8
Priority: optional
Section: interpreters
Installed-Size: 296
Maintainer: Ubuntu Core Developers <ubuntu-devel-discuss@lists.ubuntu.com>
Original-Maintainer: akira yamada <akira@debian.org>
Architecture: i386
Version: 1.8.5-4ubuntu2
Depends: libc6 (>= 2.5-0ubuntu1), libruby1.8 (>= 1.8.5)
Suggests: ruby1.8-examples, rdoc1.8, ri1.8
Filename: pool/main/r/ruby1.8/ruby1.8_1.8.5-4ubuntu2_i386.deb
Size: 217212
MD5sum: d2cedcce835b84af193332a413344b89
SHA1: 40c93e9a5df7ea9ed2567f6c8d843282c69a800a
SHA256: 7e756450dbdc504f94f67d936594c28d48923557fb7d4c32d6e9bfe147123729
Description: Interpreter of object-oriented scripting language Ruby 1.8
Ruby is the interpreted scripting language for quick and easy
object-oriented programming.  It has many features to process text
files and to do system management tasks (as in perl).  It is simple,
straight-forward, and extensible.
.
This package provides version 1.8 series of Ruby.
.
On Debian, Ruby 1.8 is provided as separate packages.  You can get
full Ruby 1.8 distribution by installing following packages.
.
 ruby1.8 ruby1.8-dev ri1.8 rdoc1.8 irb1.8 ruby1.8-elisp
 ruby1.8-examples libdbm-ruby1.8 libgdbm-ruby1.8 libtcltk-ruby1.8
 libopenssl-ruby1.8 libreadline-ruby1.8
Bugs: mailto:ubuntu-users@lists.ubuntu.com
Origin: Ubuntu
Task: kubuntu-desktop

As you can see the full install of ruby and it's tools would be:

sudo apt-get install ruby1.8 ruby1.8-dev ri1.8 rdoc1.8 irb1.8 \
ruby1.8-elisp ruby1.8-examples libdbm-ruby1.8 libgdbm-ruby1.8 \
libtcltk-ruby1.8 libopenssl-ruby1.8 libreadline-ruby1.8

So far so good. Maybe you need to add the the universe repositories to your /etc/apt/sources.list to install all the packages.
It would be nice if debian could provide a meta-package for those people who want all the ruby tools anyway.

Installing ruby on rails on ubuntu 6.06 LTS by compiling nearly anything

2007-05-02T15:11:00.000+02:00

#!/bin/sh

sudo apt-get install build-essential

cd /tmp

# download
wget ftp://ftp.ruby-lang.org/pub/ruby/1.8/ruby-1.8.6.tar.bz2
wget http://rubyforge.org/frs/download.php/17190/rubygems-0.9.2.tgz
wget http://www.zlib.net/zlib-1.2.3.tar.gz
wget http://www.sqlite.org/sqlite-3.3.17.tar.gz

# unpack
tar xf ruby-1.8.6.tar.bz2
tar xf rubygems-0.9.2.tgz
tar xf zlib-1.2.3.tar.gz
tar xf sqlite-3.3.17.tar.gz

# install
sudo apt-get install libreadline5 libreadline5-dev

cd /tmp/sqlite-3.3.17 &&amp; ./configure && make && sudo make install
cd /tmp/zlib-1.2.3    &&amp; ./configure && make && sudo make install
cd /tmp/ruby-1.8.6    &&amp; ./configure && make && sudo make install

cd /tmp/rubygems-0.9.2 && sudo ruby setup.rb
sudo gem install sqlite3-ruby -y
sudo gem install rails -y
sudo gem install mongrel -y

# create a rails test application
cd /tmp
rails rails_test --database=sqlite3
cd rails_test/
./script/generate model user name:string
rake db:migrate
./script/generate scaffold user
./script/server

If you need more you can look at the great deprec tools.
Update: A colleague told me to use "checkinstall". So we will not bypass the package management. Good point.

locate a line of code

2007-04-30T10:38:00.000+02:00

Sometimes you want to know in which file and line a method is defined.
This snippet should do that job.

Favourite Firefox Plugins

2007-04-29T10:37:00.000+02:00

http://kb.mozillazine.org/Command_line_arguments

Really simple templates with ruby

2007-04-26T12:00:00.000+02:00

Example:

template = 'Hello #{name}!'
name = 'Paul'
puts eval(%{"#{template}"}) # => Hello Paul!
name = 'Holger'
puts eval(%{"#{template}"}) # => Hello Holger!

Note that the template definition use a single qutoed string, so no variables are evaluated within the template string.
Also note that the template can't use unescaped double quotes, because the template evaluation uses them to delimit the template string.

Positives Denken

2007-04-26T11:12:00.000+02:00

...korrekt, korrekt, jefällt ma, jeht ab!