18

I wrote an important test-case for this complicated/hard-to-understand program and i can proudly say "it doesn't work on my machine". On other machines it works as expected.

Should i really investigate this or focus on other tasks?

Comments
  • 5
    Interesting issue you have here buddy, If I were you I would try clearing the cache, restart server/container/image, clean everything and try again with a fresh copy of code.
  • 2
    @rising i already cleared the files created/used by the program, no difference. The test-case calls the program 3 times and compares the output, it should be the same. But so many things can tick the testcase off, e.g. iterating over a set (python). I simply ask the guys about their environment/results tomorrow
  • 3
    A lot of things can go wrong during testing: interdependent tests that mess with the state and variable values, global settings not being cleared before ruining each test.
    I once had a problem running some tests, they were unit tests and I made sure there are no dependencies between them, still they were failing if ran all together one after one, so the only solution I found back then was to run each test in process alone. And they passed.
  • 2
    @rising i thought of this too, that's why i used the subprocess-module and wait some time until i run the program again. Therefore each process should run independently... should...
    Maybe the test works and the program is wrong, this wouldn't be the first time
  • 2
    @YouAreAPIRate you need to figure first what's the incorrect one, is it the test or the code you're testing. That's why I like TDD.
  • 2
    @rising i'm pretty sure it's the code. Because i inspected the output files and found the difference where the testcase told me to
  • 2
    But why is it working on other machines?
  • 2
    @rising maaaaaagiiiic. Seriously, no idea. I even made sure to work on the latest commit. I'll ask for the testcase output tomorrow. That still leaves the question why that testcase is working for others. But if the code is the problem i should work on other things. If my changes break the code it's not as bad because the code already was broken
  • 2
    @YouAreAPIRate you know what, just move on.
  • 2
    @rising seems like you're just as annoyed by this as i am. Sorry for dragging you into this ^^"
  • 0
    Often it is because other people run either older or newer builds of frameworks or environments, had that countless times where my node was outdated, but other people and my server just recently had fetched it, so they were of course up to date.
  • 1
    @YouAreAPIRate keep us updated
Add Comment