I’m writing a small suite of scripts for my set of Epic apps I will be testing. Taking a cue from my past experience with SilkTest and Eggplant docs related to code reuse, I’m going to Visio out the architecture. Figure I would have a “zero state” set of scripts that all apps can use that do things like open Hyperspace, login a specific Model user, etc. Then, a collection of scripts that do parts of workflow, reusing common steps where possible and calling those from a functional or regressions script. While I’m still getting used to the Eggplant implementation we have and figuring out where I can do some things and where other things need modules like Manager to work, at least the overall design can still be defined.
Any recommendations in this group from people who might have already gone through this exercise to model their test code similar to a C program where global variables are stored in “header” scripts, common snippets stored in a shared script collection and then the main body of functional scripts are organized by workflow?
Hi Chris. I have a script called CommandDepot that does nothing but house functions and handlers. I use dot notation (e.g. run CommandDeport.SetSearchRectangleCommand) to call those handlers. I then create Suite Snippets that make it easier for me to remember which parameters each of these handlers require. Each step or handoff in a manual script becomes a separate ePF script (e.g. HyperspaceStart, Login, AddAllergy), and I stack those scripts in sequence in a .txt file. I then have a shell scripts that calls each line of the text file in sequence.
1 Like