Software Configuration Management

One of the things that frightens anyone who develops code the most, or more accurately the people who are responsible for releasing or deploying that code, is the possibility that you might not have access to it within your software configuration management tool or version control system when you need it. Computers are fickle things; unless you are prepared to spend fantastic amounts of money for a truly redundant hardware/software infrastructure, there is always the chance that your system could be inaccessible when you need it most. And this doesn’t even begin to address potential network outages.

AccuRev, because of the ease of backup and restore for one thing, and the always accessible locally resident Workspaces for another, is a great insurance policy against data loss and brings developer down-time to zero. But let’s say that one of those dreaded network outages or hard disk failures occurs exactly at the time when you need to pull the latest source over to your build system for a critical release. Or imagine any kind of similar roadblock you’ve run up against in your own environment. This is yet another area where AccuRev, with just a tiny bit of prep and foresight, can save you hours, if not days, of aggravation by ensuring that you always have access to your latest “buildable” code.

How does this happen? Well, consider first the AccuRev stream paradigm where - unlike branches - there is a well defined progression of code, from the high instability of a developer’s Workspace to the “always buildable” top-level release stream. You simply have to choose the stream that you would consider as your ideal location to retrieve “offline” code from. In the following StreamBrowser, I’ve chosen the top-level “Claims_Client” stream.

Create an AccuRev reference tree based on your chosen stream. Define the location on disk where that reference tree should populate the code to. Then, use a trigger (we provide examples with exactly this functionality) to Update that reference tree anytime a Promotion happens into this stream. The end result is that you have a specific file-system location that is *guaranteed* to have the latest and greatest code from any given stream that you want. Also, this approach is optimized because it only has to transfer changes, not the entire source tree, and there is zero manual intervention required! So in the case where the network or SCM server goes down, no problem for you; your code is happily hanging out right where you need it to be.

Some organizations are much more tolerant of this kind of scenario than others. If so, can you think of other times or examples where this kind of functionality would be helpful to you?