This article specifies coding standards and tools that can be used to create reliable scripts.
It is abstracted from a longer paper.
You can either download the full Word document by clicking on the link at the bottom of this page,
or view it in Microsoft Word.
Why Scripts Matter
Kintana Deliver provides an elegant mechanism for managing the process flow of software as you migrate from development, through quality assurance to the live environment.
The product also enables you to execute custom scripts whenever you execute a command step, typically at the transition points between environments.
How reliable the process is depends absolutely on how good your scripts are at checking and warning you when errors occur.
When a step does fail , how you write the script can make the difference between quickly diagnosing the cause of the problem and moving on, and getting into a long debugging session.
Check List
Have you?
 | Added a standard # header to your script? |
 | Printed (all) the parameters used and their values on entry to the script? |
 | Used the template cl_comments to display the server on connection and disconnection? |
 | Used cl_comment not # for in-line comments? |
 | Put code block white-space separators in the log for readability? |
 | Identified commands that return success (0) exit codes even when they fail and used sscmdshellNT to handle them? |
 | Identified commands that return error (non-zero) codes when they have actually succeeded and handle them ( requires sscmdshellNT filter logic upgrade)? |
 | Error trapped batch files run under the NT Command Shell and tested the exit code in the Bash script as per standard? |
 | Simulated and tested failure condiitions for all critical points in the script? |
|