A really sly option would be to, effectively, allow infinitesimals. So 10/0 would return 10/0 (perhaps a kind of NaN but with more information retained), and if you multiplied it by 0 elsewhere you'd get 10. Probably a bad idea for novice programmers but interesting to consider.
How to handle division by zero in a language that doesn't support exceptions?
21–30 of 66 posts
Re: How to handle division by zero in a language that doesn't support exceptions?
#22A really sly option would be to, effectively, allow infinitesimals. So 10/0 would return 10/0 (perhaps a kind of NaN but with more information retained), and if you multiplied it by 0 elsewhere you'd get 10. Probably a bad idea for novice programmers but interesting to consider.
Worse yet, just return the highest possible value of the data type. Probably a 32bit integer.
Re: How to handle division by zero in a language that doesn't support exceptions?
#23Sounds like a new PHP in the making, someone even mentioned it in the link. I think it's the only language I have worked with that's so happy to chug along and ignore errors that arise in execution. And the exception system was just thrown on top of it in PHP5 as an afterthought. Ugh.
Visual Basic has this mode. See 'on error resume next':
http://msdn.microsoft.com/en-us/library/aa266173(v=vs.60).as...
Re: How to handle division by zero in a language that doesn't support exceptions?
#24I liked the ternary operator where you provide an alternate value in case the denumerator is 0.
One more idea I didn't see is to mark the variable as "invalid". A general "this variable doesn't have a value because the program did something bad", which carries a full stack trace with itself. Then, if you ever use that value anywhere, the result is again invalid, with the new operation attempted on top (i.e. if the invalid value ever appears in an expression, the entire expression yields invalid and no function calls are attempted; an exception would be boolean operators - true or invalid is true and false and invalid is false - there may be other exceptions to that, too). Once that value is no longer referenced anywhere, you log a nice long description of the entire chain where things were wrong and continue business as usual. Maybe email the trace to the project lead or something.
Something like the Maybe monad basically. It will work with anything else that can fail, too, and you can allow people to construct their own invalids. And the program will continue to run as long as the error isn't too deep, which is what is really wanted here, I think.
Re: How to handle division by zero in a language that doesn't support exceptions?
#25Sounds like a new PHP in the making, someone even mentioned it in the link. I think it's the only language I have worked with that's so happy to chug along and ignore errors that arise in execution. And the exception system was just thrown on top of it in PHP5 as an afterthought. Ugh.
I think it's the only language I have worked with that's so happy to chug along and ignore errors that arise in execution
Other technologies designed to render content for web browsers do this also, such as javascript and the django template rendering engine.
Re: How to handle division by zero in a language that doesn't support exceptions?
#26For IEEE floating point, there is NaN and +/-Infinity which are raised when dividing by zero: http://grouper.ieee.org/groups/754/faq.html#exceptions For integers that don't have NaN/infinity values the options are less clear. Note that languages with two's complement fixed width signed types like int32 have the same problem with INT32_MIN / -1, since the result is not representable as an int32.
Yeah, but they usually just say that the result is undefined. At least that's what C and C++ do. And it's useful for optimizers, because they can assume more about basic arithmetic in signed types and do some neat tricks.
> At least that's what C and C++ do. And it's useful for optimizers,
And for security exploits.
Re: How to handle division by zero in a language that doesn't support exceptions?
#27I liked the ternary operator where you provide an alternate value in case the denumerator is 0. One more idea I didn't see is to mark the variable as "invalid". A general "this variable doesn't have a value because the program did something bad", which carries a full stack trace with itself. Then, if you ever use that value anywhere, the result is again invalid, with the new operation attempted on top (i.e. if the in…
> One more idea I didn't see is to mark the variable as "invalid".
That's what NaN is.
Re: How to handle division by zero in a language that doesn't support exceptions?
#28I liked the ternary operator where you provide an alternate value in case the denumerator is 0. One more idea I didn't see is to mark the variable as "invalid". A general "this variable doesn't have a value because the program did something bad", which carries a full stack trace with itself. Then, if you ever use that value anywhere, the result is again invalid, with the new operation attempted on top (i.e. if the in…
I like the idea of providing an alternate value or even alternate flow of execution.
The action to take when dividing by 0 is based on context: it is a business behavior. In my opinion, it is no different than any other business requirement (for example: passwords must be longer than 8 characters, etc.).
A language should allow a developer to configure the action to take when a division by zero occurs within a given context: the default being an exception and/or termination of program execution.
Re: How to handle division by zero in a language that doesn't support exceptions?
#29I liked the ternary operator where you provide an alternate value in case the denumerator is 0. One more idea I didn't see is to mark the variable as "invalid". A general "this variable doesn't have a value because the program did something bad", which carries a full stack trace with itself. Then, if you ever use that value anywhere, the result is again invalid, with the new operation attempted on top (i.e. if the in…
> One more idea I didn't see is to mark the variable as "invalid". That's what NaN is.
Not quite, marking a variable as invalid in this context is essentially a job for the type system. For example, in C, you can have a void* which can be null. This can lead to seg faults, and NaN has the same pitfall. Instead, with the Maybe idea above, the type system explicitly makes you handle the NaN case since the type is a discriminated union. This is like `int?` in C#.
Re: How to handle division by zero in a language that doesn't support exceptions?
#30 result, err := x / y
if err != nil {
do something
}