The mailing script will sent an e-mail to the commit log mailing list with information on the files that have been changed and the diff’s between the previous version along with more general information, like the name of the author and the timestamp of the commit.

You can subscribe to this list from the Support page at the project’s website.

Too bad they still include that deprecated Perl mailer script in the distribution. It took me some time to find out that mailer.conf was only to be used by the Python mailer script. And even then, the Perl version doesn’t support SMTP auth (which I added myself, but that doesn’t matter anymore).

The service I mentioned in a previous blog post will eventually also run from my development server since I don’t want to clog Martijn’s bandwidth with my stuff.

From now on the Subversion repository is served directly from my development server at home. I had previously enabled the Subversion repository at SourceForge for this project so Ohloh could analyze the code. But now that I have my server running non-stop I can serve the repository myself. So I disabled the repository at SourceForge again and updated the enlistment at Ohloh to query my own server for updates.

The location for the source code repository is: svn://source.irc-collective.org/

I could of course host the entire project at SourceForge, and save myself from doing all kinds of maintenance. But the point is, I take great joy in setting up all those services. I learn something new every time. Of course, it can be frustrating when things brake or when they take up much more time than expected. But once I get things working, I get a great deal of satisfaction from it.

I might want to see if I can use svnsync to synchronize the source code repositories between SourceForge and my development server, but that’s something I’ll save for later.

I still have to move ViewVC to the development server, otherwise I have to copy (dump and load) the entire repository from my server to Martijn’s server with every change.

Update 25/10/2009: The development server which hosts the source code repository is offline again. The power consumption of the server is simply too high. Until I buy a replacement with much lower power consumption it will remain offline.

• a selection box to limit the number of results per page
• and a check box to show or hide non chat messages.

logfile.pl rev41

The code for the web application is in some really bad shape. What’s killing me, is the fact that the view and the logic are not separated from each other, along with a whole bunch of other bad practices. If anything needs a rewrite, it’s going to be this part of the software!

On the same note; the entire web application could use a paintjob! A neat CSS stylesheet might just do the trick though.

“Providing solutions to archive, query and analyze IRC logfiles.”

My idea of “solutions” is twofold; On one hand, I should create software so users can archive, query and analyze IRC logfiles themselves. On the other hand, I should setup a service that provides the same functionality to users that don’t have the time or know how to install and configure the software.

At the time of writing, the software itself is not mature enough to actually start working on such a service.  But, I can ponder about it from time to time and publish some of the ideas here on this blog. Besides, I should spent that spare time on developing the actual software.

The service would basically be a website where users can register themselves and the channel they want to log. After they have registered, they can either invite a logging bot that will start logging the channel or upload their own IRC logfiles. Users should then be able to make use of the same functionality that the web application provides, such as viewing a logfile, browse the image gallery, create a paste, etc.

Back to coding…

To straiten this out, I installed mIRC on a Windows box, created a copy of the original config file, than ran mIRC and configured logging according to the instructions in appendix B. I than finally compared the new config file with the copy.

I was surprised at first to find out that mIRC doesn’t store all the changed settings as separate variables, but instead uses a set of obscured serialized variables that hold more than one setting each.

Then I  realized that for this reason Martijn (a supporter and contributor to the project) had to incorporate the instructions for mIRC as a GUI walk through, rather than listing commands that the user must type (as is the case with the other clients). A real bummer.

A nice feature of Ohloh are some of the metrics it can derive from the source code. But, before Ohloh can analyze the source code of your project, it needs access to a source code repository (CSV, SVN or Git). I was going to mirror the repository on SourceForge.net anyway, so now I had another reason to do it.

Check out the code analysis for the IRC Collective project for yourself.

Don’t forget to register yourself on Ohloh and add IRC Collective to your stack!

From the synopsis of the parselog script it read that you can set the path where local copies of images and thumbnails are stored from the command line using –imagedir. Turns out that this option was not yet implemented. The setting could be set from a config file, but not as an argument to the parselog script.

While fixing this I noticed that it’s better to install the Perl modules on the workstation as well so EPIC (plugin for Eclipse to support Perl syntax highlighting and object outline) wouldn’t yell at modules it can’t find. Haven’t made the trip to CPAN as of yet, but I’m planning to!

I tested the change and it works like a charm. Then I committed the changes to the source code repository and revision 40 was called into existence.

On to bigger challenges…

SPAM
In my battle against SPAM, I took another measure to protect the forum and bug tracker from fake registrations and bug reports by adding CAPTCHA to the forum registration page and the bug report page. This is going to save me some time, I was a little bit fed up with having to remove fake registrations and bug reports. That time is better spent on adding new features!

Latest revision
I checked my development server for the latest revision of the source code and it seems that it is a little bit more recent (rev 39) than the source that is available online (rev 36). Just some minor changes though, mostly documentation related updates. I will synchronize those soon. Speaking of synchronization, I’m thinking of using a SourceForge Subversion repository. That way it will be more accessible and centralized.

The todo list and roadmap are still in sync with the latest revision. I might as well add all the items on the todo list (and roadmap) to the bugtracker as feature requests. This will give them status and if others want to join in they can see what is assigned and what is not.

Working installation
The current installation of IRC Collective on my development server is up to date and working! It has not ‘assimilated’ any IRC logfiles for the past 15 or 16 months, but other than that it is still working!

Eclipse workspace
Setting up the Eclipse workspace was not as easy as I though it was going to be. Turns out I used NFS  to share my home directory on the development server. The Eclipse workspace was stored within this directory. It required some additional software to be installed on my workstation before I could mount the NFS share again. Luckily the workspace was compatible with the version of Eclipse I have currently installed.

Reboot complete!
That’s it! I can start developing again…

1. Masked all email addresses on the website
2. Pruned all bogus accounts and posts on the forum
3. Updated the forum and the bugtracker to the latest versions
4. Re-enabled the forum
5. Re-enabled anonymous bug reports and feature requests (bugtracker)

Most of the site is working again, except for email. But this is not a priority for now. First I want to see if spammers are still able to put posts on the forum using non activated accounts. I hope the latest version of punBB is more secure.

I just need to install some more development tools again and I can start working on the website. To be continued..