1) Select the barcode type you want to use. For common stuff, Code 39 or Code 128 are good options. Make sure the barcode scanners available support the selected barcode type.

2) Create an algorithm for encoding the information you want to embed on the barcode.

2a) It’s a good idea to use Base36 encoding to shorten the data and therefore make the barcode shorter.

2b) It’s also a good idea to use a checksum algorithm if you use Code39 (Code 128 already has a checksum algorithm), for example Luhn’s algorithm.

3) Bind the data to the barcode control. Barcode controls are usually expensive, so if you want a cheaper solution go for a barcode font, there are free ones out there.

Today while I was setting up a Subversion 1.5 installation over Apache 2.2.10 on Windows, I was experiencing a strange difficulty. After making the necessary configuration details on httpd.conf and I tried running the server, everything went smoothly, however when I made it run as a service, it would not run and give the following error:

“Cannot load C:/../modules/mod_dav_svn.so into server: The specified module could not be found.”

Even though this module was at the right location. After googling for a while and reading the Subversion FAQ for this problem and another useful document, I tried all the suggestions but the problem would still persist. After a while of tinkering, I resolved it by copying the following libraries from the Subversion’s installation to the modules folder of Apache:

• intl3_svn.dll
• libdb44.dll
• libeay32.dll
• libsvn_client-1.dll
• libsvn_delta-1.dll
• libsvn_diff-1.dll
• libsvn_fs-1.dll
• libsvn_ra-1.dll
• libsvn_repos-1.dll
• libsvn_subr-1.dll
• libsvn_wc-1.dll
• ssleay32.dll

I hope this post helps those who are having problems with Subversion deployments.

It did not use to be like this in previous Subversion releases, usually after putting the *.so files on the modules entry on Apache and configuring httpd.conf it was enough to make it run…

What’s Continuous Integration?

As Martin Fowler said, Continuous Integration is a software development practice where several people integrate their work frequently, usually each person integrates at least daily – leading to multiple integrations per day.

One specific point of Continuous Integration is that the verification of the integration is done automatically using what is called a Continuous Integration Server. This server will fetch the latest code from the corresponding VCS in use (hopefully Subversion, read my article about it).

Why Continuous Integration?

The most important advantage of following the Continuous Integration practices is that errors are detected in a much earlier phase, specially when you use it along with Test Driven Development. Errors in this context not only mean bugs but integration errors as well, as integration is done on a daily basis.

It is not the only advantage of course, for example, deployments to production are a LOT easier as the build process is already in place. I recommend reading the benefits section of Fowler’s article for a detailed explanation.

Tools and Useful Resources

I have used the following tools for doing Continuous Integration:

• CruiseControl
• CruiseControl.NET

I also heard good things about Hudson, so give it a try as well.

Useful resources:

Martin Fowler on Continuous Integration

6 Steps to Successful Continuous Integration

Enjoy!