Issue15815

Created on **2012-08-29 19:24** by **terry.reedy**, last changed **2012-09-26 19:44** by **terry.reedy**. This issue is now **closed**.

Messages (8) | |||
---|---|---|---|

msg169410 - (view) | Author: Terry J. Reedy (terry.reedy) * | Date: 2012-08-29 19:24 | |

I propose that we add 'of ####' to all ZeroDivisionError messagesso it is less necessary to use a debugger to add print statements to recover the information currently tossed away. (I also propose to change denominator from 'zero' to '0', '0.0', '0+0j', as appropriate.) >>> 3//0 ... ZeroDivisionError: integer division or modulo by zero # augment to ZeroDivisionError: integer division or modulo of 3 by 0 Similarly for >>> 3 / 0 ZeroDivisionError: division by zero # perhaps this should be 'true division of 3 by 0' >>> 3.0 / 0.0 ZeroDivisionError: float division by zero >>> (3+3j) / (0+0j) ZeroDivisionError: complex division by zero In #7482 it was proposed to make float and complex messages same as int message by adding 'or modulo'. I an not sure why not to do that, or if it was just deferred. Fractions currently print the numerator as part of an invalid numerator / 0 result from *either* construction or division. >>> from fractions import Fraction as F >>> F(3, 0) Traceback (most recent call last): File "<pyshell#20>", line 1, in <module> F(3, 0) File "C:\Programs\Python33\lib\fractions.py", line 167, in __new__ raise ZeroDivisionError('Fraction(%s, 0)' % numerator) ZeroDivisionError: Fraction(3, 0) # Here, 'Fraction(3, 0)' is invalid input. >>> F(3, 1) / F(0, 1) Traceback (most recent call last): File "<pyshell#19>", line 1, in <module> F(3, 1) / F(0, 1) File "C:\Programs\Python33\lib\fractions.py", line 367, in forward return monomorphic_operator(a, b) File "C:\Programs\Python33\lib\fractions.py", line 417, in _div a.denominator * b.numerator) File "C:\Programs\Python33\lib\fractions.py", line 167, in __new__ raise ZeroDivisionError('Fraction(%s, 0)' % numerator) ZeroDivisionError: Fraction(3, 0) # Here, 'Fraction(3, 0)' is invalid output. I found this confusing until I worked out the dual meaning. I think ZeroDevisionError: invalid Fraction(3, 0) from construction or division might be a bit clearer. I have not looked at decimals. |
|||

msg169418 - (view) | Author: Antoine Pitrou (pitrou) * | Date: 2012-08-29 21:15 | |

What would be the point? Usually you should find out why the denominator is zero, not what the numerator's value is. |
|||

msg169419 - (view) | Author: Terry J. Reedy (terry.reedy) * | Date: 2012-08-29 21:32 | |

In case the value of the numerator helps find out why the denominator is 0. The example given by Mike Graham on python-ideas, Verbose traceback formatting was def f(a): x = a * 4 y = a - 4 return x / y |
|||

msg169420 - (view) | Author: Antoine Pitrou (pitrou) * | Date: 2012-08-29 21:35 | |

> In case the value of the numerator helps find out why the denominator > is 0. The example given by Mike Graham on python-ideas, Verbose > traceback formatting was > > def f(a): > x = a * 4 > y = a - 4 > return x / y Well, in this case y == 0 iff a == 4. I don't see how the numerator helps at all. (and besides, it's a rather silly artificial example) |
|||

msg169556 - (view) | Author: Ezio Melotti (ezio.melotti) * | Date: 2012-08-31 16:42 | |

I'm not sure this is really useful. Once I know that I divided by zero and I know where the division is, I don't really care what the numerator is. Knowing the exact type of the denominator (0, 0.0, 0j) doesn't sound too useful as well -- unless you somehow manage to get the wrong type and a 0-value at the same time. On a side note, I don't like ZeroDivisionError: integer division or modulo of 3 by 0 too much, but I would prefer ZeroDivisionError: integer division (or modulo) by zero over ZeroDivisionError: integer division or modulo by zero IIRC the reason why both "division" and "modulo" are included in the message is because the error is created in a piece of code that is shared by both operations. |
|||

msg171354 - (view) | Author: Ezio Melotti (ezio.melotti) * | Date: 2012-09-26 18:57 | |

Terry, do you still think that adding the numerator would be useful? |
|||

msg171355 - (view) | Author: Mark Dickinson (mark.dickinson) * | Date: 2012-09-26 19:11 | |

I'll add my -1 to this; I don't really see the use. |
|||

msg171358 - (view) | Author: Terry J. Reedy (terry.reedy) * | Date: 2012-09-26 19:44 | |

Not enough, at present, to leave this open. |

History | |||
---|---|---|---|

Date | User | Action | Args |

2012-09-26 19:44:32 | terry.reedy | set | status: open -> closed resolution: rejected messages: + msg171358 |

2012-09-26 19:11:09 | mark.dickinson | set | messages: + msg171355 |

2012-09-26 18:57:33 | ezio.melotti | set | messages: + msg171354 |

2012-08-31 16:42:52 | ezio.melotti | set | messages: + msg169556 |

2012-08-29 21:35:32 | pitrou | set | messages: + msg169420 |

2012-08-29 21:32:21 | terry.reedy | set | messages: + msg169419 |

2012-08-29 21:15:08 | pitrou | set | nosy:
+ pitrou messages: + msg169418 |

2012-08-29 19:24:44 | terry.reedy | create |