Sunday, January 15, 2006

ClearCase Nightmares

For those who don’t know, ClearCase is the IBM solution for source code management.  I am using it on a project for a client and it has been a nightmare.  It regularly corrupts source files and resource files.  Getting everything working again can be challenging to say the least.  I think the problem is the way ClearCase views files as a stream of characters.  It has no concept of methods, properties, or other characteristics found in modern source code.  That makes merges very interesting.  It seems to get confused very easily and when that happens, what you get out of it can be just as confusing.

It seems to follow this algorithm: Take random bytes from the file in the repository, and merge them in random and creative ways with random bytes from the corresponding file on your local drive.  Repeat this process until the end of both files is reached to produce the final merged file for use in your project.  As you can imagine, this produces some very interesting results that make the compiler frown and grumble about various things and puts you on the path of a day-long journey to sort out the whole mess.  I wonder if IBM has a QA group for the software development teams…

I’ve used several SCM products on various projects and they all seem to have their quirks and gotchas.  I haven’t had that warm, fuzzy feeling of knowing my files are safe and secure, and what I put in will come back out in the same condition when I make changes.  So, to end my rant, let me make a prediction.  If Microsoft’s Team System contains a good, friendly, reliable, stable, source management system (and I haven’t used it, so I don’t know) it will dominate the market for source control.  I hope they got it right, and I hope I get a chance to use it soon on a project.

0 Comments:

Post a Comment

<< Home