This weeks biggest time waste bug

What is this weeks longest to figure out bug? Almost 2 days wasted.

As someone who is not that familiar with unix, I got stuck on some testing this week. Basically we are developing a new version of our app, and the installer is a bit tricky, and on linux, there is no uninstaller yet, you have to manually run a script to uninstall. So not knowing how to uninstall, I just re-installed, the installer takes a commandline flag which it uses to register with the central server, and then starts a background service. And since running the installer when given the flag registered with the backend correctly, I assumed it’s working perfectly. NOT.

Turns out the installer does not overwrite prior settings, so while the installer phones-home, the background service never got reconfigured. Serves me right for assuming things. Also need to write more notes on how to do the arcane steps on linux.

1 Like

Don’t know about this week’s, but my record for a bug that took the longest time to track down was a UI issue in a previous role which fell into the category of “Did I just see it do X?”

It was a page in the middle of a workflow process which had a start button at top left, a ‘next’ button at bottom right and a pair of time selection spinners roughly in the middle of the page. The spinners were there to set a service call time. I constantly had a nagging feeling that the times were out by either an hour or a minute on some (but not all) processed appointments. Eventually I realised that when stepping through the workflow, when I was completing that page, I was moving the mouse up to the top of the page in preparation to click the start button on the next page, and if I happened to move the wheel on the mouse during that move, and the pointer passed through the spinners whilst the next page was still loading, I was putting an input into the time selection spinner as the pointer passed through them on my way back to the top of the page.

It didn’t happen every time, it didn’t always trigger the spinners moving on by one minute or one hour. But by constant testing of the workflow, I’d gotten pretty slick at using the app and so my mouse movements were increasingly fast and accurate in terms of placing the screen pointer for the next page/step of the workflow. And because I was so fast with the mouse movements, it was hard to spot what was actually happening.

It took me three months to pin that bug down.