The Zeitgeist

About 2 years ago I decided that CVS was old and decrepit enough to be just too hard to use, too brittle and I decided to get with the times and migrate all my various source archives to what all the cool kids were using, Subversion, or “svn.” And for about 2 years it’s worked out pretty well. SVN is not quite as easy to use as CVS, it’s got a lot of different access modes and methods, and it doesn’t keep your files as a set of diffs; however lately I’ve run into little unexplainable problems that have not only eaten dozens of hours of development time, but made me even question the safety of SVN itself.

SVN seems, in my experience at least, to have problems with network check-ins. I run my SVN server in a machine in one of my co-lo racks; I do this so that if my house if blown away by a hurricane, my code will still be there, also so that I can access my code from the road when I travel, or even from Startbucks if I just want to get out of the house for the afternoon. Several times in the last few months I’ve run into situations where I have checked files into SVN only to find them 1/2 checked in, or to find a file reported as “damaged” only when I try to make a new branch of my code base. YiKeS!

Not amount of SVN ju-jit-su seems to reset these files or unlock them. I have been forced on several occasions to get a good working copy, and create a new SVN repository. Needless to say, this is a problem. So, what to do? I’ve updated my Subversion code to the latest stable version, all my clients are up to date (I usually use IntelliJ IDEA, and it’s client is up to date)… yet SVN is starting to scare me.

I know nothing about GiT or Mercurial (?) or any of these other new SCMs…. worth the effort? Or am I asking for more trouble and lost development time…?

Trackback URI | Comments RSS