AlephZarro

What to say. I made Firefox 3 compatibility update of XPather. Go, and get it from XPather’s site or soon also from AMO. Bye.

Oh yes, since FF 3 beta 4, the DOMinspector is not anymore distributed with FF. You need to install it as a separate extension.

Oh, XPather has always been just a small development-utility toy-project, though it helps (a lot?). Anyway, the facts are inspiring: 2,5 year old; 58000 downloads; very positive feedback; $7 donations (+5 empty promises); $27 adsense. The mission: contribution to public successful. Isn’t it? ;)

You do Ruby. You are on *nix. You have a task you need to schedule using cron from your scripts. Later you schedule another task. After some time, and probably from another script, you want to unschedule/cancel the first task … How do you do it?

Fetch CronEdit …

    sudo gem install cronedit

….and just do it:

    Crontab.Add  'agent1', '5,35 0-23/2 * * * echo agent1'
    ...
    Crontab.Add  'agent2', {:minute=>5, :command=>'echo 42'}
    ...
    Crontab.Remove 'agent1'

You can CRUD your tasks independently (based on the IDs you give them). The schedule can be in Hash notation, or the standard cron text definition.

UPDATE: The new version of library allows you to do some more tricks. First of all, you can manipulate groups of definitions (bulk/group operations). Besides modifications of the actual crontab, you can do now all operations upon files or in memory. And you can combine these ‘devices’ as needed. Let me show you a small example:

    fc = FileCrontab.new '/jobs/mail-delivery.cron'
    Crontab.Merge fc
    ...
    Crontab.Subtract fc

The previous snippet loads a group of cron definitions from file and merges them with the current system crontab. On the last line the same group of definitions is removed from the system crontab.

See documentation for more details and examples.
UPDATE: A related article discusses the library in context of scheduling of Ruby Agents using Cron.

In Ruby there are few schedulers out there, for example the most advanced: openwferu scheduler. They are great for many task, yet…. they are Ruby proprietary schedulers and the host ruby script needs to be running. If you need native cron scheduling, for whatever reason, go for CronEdit.

Do you want to have photo albums (on your site)? I did. I wanted a comfortable but not “over-web20-ized” user experience, and above all, simple and pragmatic content management/control.

Believe or not, nothing I came across was just the right thing. Too heavy, too weak, too ajaxy, unacceptable license, etc. At the end of the day I found out I had not had to look too far. Ondrej Jaura (a good friend of mine) created a simple command-line album generator - Rhotoalbum. Bonus: it is written in ruby :)
You treat your images as ordinary files in directories (which stand for albums) and run a generation command. That’s all you need to do. You add new images or change something? Just run the same script - anytime. It creates necessary thumbnails, and simple style-able index pages. For slideshow purposes it (optionally) integrates with nice TripTracker.

My way of appreciation of my Ondrej’s great work was … to contribute the project and help out in making it better. This resulted into a release of a new stable version 0.3 these days.

Well, I started this post saying I wanted some place for our photos to show them to friends occasionally. Here is are our Rhoto photo gallery. Another showcase for Rhotoalbum is Ondrej’s photo gallery.

In Summer 2005 I started few home-made experiments with workflows for governing complex processes - such as system integration on Web layer. I did this as a part of my PhD activities. The main motivation had been my deep disappointment with all the “enterprisy” workflow systems bull**** laying around. I needed something very dynamic, light-weight and portable. I needed a hammer !

I engaged with Mozilla platform and built a workflow engine on top of this unbelievably feature-rich and mature open-source technology. This platform, surprisingly, allowed for much more. So I added libraries for Web navigation, information extraction, rich GUI, etc.

At the end of the day, there it was: a system for development and execution of full-fledged personalized on-top-Web applications. Users could create and share their own scenarios, own personal agents for searching best cars, checking back account, performing complex transactions on portals, or making any mashups crimes. You can learn more about it on the LumberJaczk homepage. From the very beginning the system was meant to be open-source as it really fits the model.

In mids 2006 a wave of strong interest arouse from commercial segments. Unfortunately this interest clashed with the open-source strategy.

May 2007 - The end of the story (or beginning of another one). Lixto Software GmbH acquires the LumberJaczk technologies. Lixto is at the moment one of the market-leaders in the field of Information Extraction and a solution provider for Online Market Intelligence, Web Process Integration and Metasearch. Openly, the current plan of the company is to keep the source closed.

I hope interesting things will happen!

It goes like this: Firewatir, the Firefox port of WATIR, depends on JSSh. And JSSh does not install & run nicely on Linux. Clear ?
Annoyed by this situation, I looked what’s the problem with this JSSh. To make the story short, success - downlaod the working JSSh for Linux here.
Update 2007-04-11: modified to work also with FF 1.5.3+

Oh, by the way, JSSh stands for JavaScript Shell and it allows you remotely manipulate Firefox remotely. Firewatir gives you a higher-level toolkit for “Web Application Testing in Ruby” (of course not Web scraping ;>>> ). Talking about it let me show how to install the whole thing:

(0. I would recommand to create a dedicated Firefox profile run Firefox in a dedicated profile firefox -ProfileManager)

1. Downlaod and install JSSh extension to Firefox and restart it.

2. Start Firefox:
firefox -P testyard -jssh

The JSSh should be up, listening on port 9997 (note: consider security issues). You should be able to communicate with Firefox via telnet. For example:

>telnet localhost 9997
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.
Welcome to the Mozilla JavaScript Shell!
> 1+1
2
> help()
....
> exit()

3. Dowload firewatir (you have to do it from its homepage, it is not on rubyforge yet). Use the mswin32 version, although you are on Linux. Install:
sudo gem install firewatir-1.0.1-mswin32.gem

3. Get to the real ruby test. For example:

require 'rubygems'
require 'firewatir'
include FireWatir

ff=Firefox.new
ff.goto("http://www.google.com/ncr")
ff.text_field(:name,"q").set("ruby")
ff.button(:value,"Google Search").click
puts ff.element_by_xpath("//a[@class='l']").text
ff.close

PS: For those interested in what did I fix: I took one of the latest builds from Dave, repackaged it so that it installs “properly” - as an extension of “21st century” - and added the config GUI shipped originally with the JSSh. It was tested with FF 2.0.0.1 on Ubuntu (Eft), Kubuntu, Gentoo. Update: tested also on FF 1.5(.0.7)