As a QA person I've been told by developers a myriad of rather amusing things or funny conversations caused by failed tests and the all too often "works on my machine".
I've been meaning to compile them, but I'll just start making posts as they happen/I remember.
All names and locations have been removed for privacy of those involved.
$Dev: On $ticket you mention that the $behaviour happens when it shouldn't. However on the requirements it says $behaviour is expected.$Me: Oh, I must have missed that, where does it say so on requirements?$Dev: Well, something things are common knowledge and shouldn't need to be written anywhere.
$QA2: This doesn't work$Dev: Of course it works, look *press button on his laptop* it works, you're testing it wrong.$QA2: But that's your machine.$Dev: So? It works.$QA2: "Works on my machine"$Dev: It works!$Me: Great, $Dev we should tell $Boss to arrange $Ops to have your laptop sent to $location.$Dev: Why?$Me: Your machine will be the new prod server obviously, since it only works on your machine.
At this point $Dev helped $QA2 to find the root cause of the problem, things were indeed broken.
$Me: This doesn't work when you use $resource_type.$Dev: Can't reproduce, did what arguments did you use? Can you retest please in the RC and master branch please?$Me: I used the arguments of a $resource_type, here's the error stack trace on both branches.$Dev: I used $other_resource_type and it works, are you putting the arguments correctly? Or are you using dummy/broken data?$Me: I used $resource_type, if I had used bad data I'd have said so in the first place. Just try $resource_type.*couple minutes go by*$Dev: Ok this doesn't work, I'll fix it.