In the comments of this post, Kym pointed out that I had been remiss in providing instructions, so I’ve now corrected that state of affairs. I’ve uploaded v1.1 to SourceForge, and the zip file includes an MSI installer as well as PDF and DOCX formats of instructions.
As I alluded to in the comments of my previous post, I got a little carried away once I started on the instructions. I made a few minor modifications to the logic for generating reports, and added TRX as an additional, acceptable extension. That is the default extension for MS Test results files, but during the course of my build process, the files get renamed to .xml, so I didn’t have the .trx whitelisted in my validation routines. But now, I do.
The biggest change, though, is the addition of an interactive GUI, pictured here from SourceForge:
As I was adding instructions, it occurred to me that most people wouldn’t be following my build process, so I added a WPF project to my existing solution and set about making a GUI for this over the last week in my spare time. My over-engineered, layered implementation actually came in handy because this required very little in the way of changes to the underlying architecture. I just had to add logic to it for finding solutions and test files on disk, and the rest was sufficient.
The GUI isn’t perfect, but it’s pretty functional. If people like it, I’ll be happy to add to it (or you’re welcome to snag the source and contribute yourself if you want–it’s still available from the subversion repository). The gist is that you pick a solution and the tool will show you test runs for that solution. You then choose a test run and generate a report for it.
As I became my first user, I decided that it’d be interesting to have this open while developing, and that prompted me to add a “refresh” button so that I could do a test run, refresh, and generate a report for my latest run. In this fashion, I actually spent a few days developing where I’d view all my test results with this tool instead of the Visual Studio “Test Results” window.
There is plenty of room for extension, which I’ll do if people are interested. I could poll the filesystem for new tests instead of a manual refresh, for instance. But incremental baby steps. This is one of a bunch of pet projects that I have, so it’s not exactly as if I’m flush with time to get crazy with it. I’ll let the response level guide how much work I put in.
In that vein, if you like it or even if you don’t like it, but want me to do something to make you like it, send and email or post a comment here or on SourceForge! Aside from monitoring the SourceForge download content (and who knows what that really says about the project), that’s my only way of getting feedback.