On 17/11/13 13:33, Roy Smith wrote:
Every once in a while, I'll get into a situation where something is
happening that I just can't understand. If a given pice of code is
being called, there's NO WAY the program should be exhibiting the
behavior it's exhibiting. But, there's also NO WAY that piece of code
can't be getting called. So, I stick "assert 0" in the code an re-run
the program to see if I get an AssertionError. If I do, then I know
the code is being run. If I don't then I know it's not. Either way, I
know more about what's going on than I did before. Once I know what's
going on, I remove the assert.
Every once in a while, I'll get into a situation where something is
happening that I just can't understand. If a given pice of code is
being called, there's NO WAY the program should be exhibiting the
behavior it's exhibiting. But, there's also NO WAY that piece of code
can't be getting called. So, I stick "assert 0" in the code an re-run
the program to see if I get an AssertionError. If I do, then I know
the code is being run. If I don't then I know it's not. Either way, I
know more about what's going on than I did before. Once I know what's
going on, I remove the assert.
Are assertions the right thing in that case? I'm becoming more and more
inclined to the idea that if you want to know whether code is getting
run, you should put a debug log in, and leave it in. That way it's
easier to track down the next bug (I don't know about others, but when I
write buggy code I go all-out and put lots of bugs in rather than just one).
Rob
