Division operation #18
Conversation
|
ok, i will fix it. |
|
divided by zero will get into infinite loop @MahmoudRizk |
|
@titansnow , Thanks, i will fix it. |
|
@MahmoudRizk , I found not only zero could cause this. Any number divided by nonpositive will also get into infinite loop. See this #include"bigint.h"
using namespace Dodecahedron;
int main()
{Bigint a=10;
Bigint b=-2;
a/b;
return 0;}May you forget to check if number is not positive? |
MahmoudRizk
force-pushed
the
MahmoudRizk
branch
from
9949376
to
5f160ac
Compare
…y is correct in case of signed numbers division'
|
Division by zero bug is fixed. |
|
:) |
|
I dont like the current solution to dividing by zero. Throwing an exception might be acceptable, so the behaviour is the same as with int type. |
|
changed the way of error handling, you can check it. @kasparsklavins |
|
I didn't notice that you are the one who made the subtraction fixing 😆 , forgive me if you felt insult for mentioning you by someone. @titansnow |
| } | ||
| catch(const char* msg){ | ||
| std::cerr << msg << std::endl; | ||
| std::terminate(); |
There was a problem hiding this comment.
If an exception is thrown, is it user catchable?
There was a problem hiding this comment.
no, user won't catch it. It is caught in the division function interface only and printed on screen as shown above.
It is a little tricky and i don't know how to handle the exception, so please can you clarify more.
There was a problem hiding this comment.
Throwing the error is enough, you dont have to catch it.
If division by zero occurs, let the user deal with it.
There was a problem hiding this comment.
My solution is to make an inheriting class to std::runtime_error named arithmetic_error and throw it with "Divide by zero" when divided by zero just like Java does. Like this function for checking:
void check_divisor(Bigint const &b)
{
if (b == 0)
throw Bigint::arithmetic_error("Divide by zero");
}There was a problem hiding this comment.
@titansnow I agree. That would be better.
There was a problem hiding this comment.
Ok, i will try it.
src/bigint.cpp
Outdated
| Bigint zero("0"); | ||
|
|
||
| if(q==zero || q == 0){ | ||
| //std::cout << "Error: Dividing by zero" << std::endl; |
There was a problem hiding this comment.
Remove the commented out code.
| private: | ||
| std::vector<Bigint> divide(Bigint q); // returns quotient(index[0]) and remainder(index[1]). | ||
| public: | ||
| std::vector<Bigint> operator/(Bigint q); // interface for divide() function. |
There was a problem hiding this comment.
Why does division return a vector?
There was a problem hiding this comment.
because it returns both quotient and remainder.
There was a problem hiding this comment.
divide() can stay as is, but / operator should return just Bigint so syntax like this would work a * b / c.
There was a problem hiding this comment.
I think there can be both operator/ and operator% wrapper divide and split the results, return what is needed, just like killing two birds with one stone. By the way, I think you algorithm is so cool 👍 @MahmoudRizk
There was a problem hiding this comment.
Yes make it two separate operators.
Also, could you also describe how the algorithm works and why it works?
There was a problem hiding this comment.
Ok, i will separate it.
About the Algorithm it is called "Double division algorithm", you can check this link http://www.doubledivision.org/
There was a problem hiding this comment.
Thanks for the link. That is an awesome algorithm indeed.
|
|
||
| if(sub_p<look_up[i] && i!=0){ | ||
| tmpx1=look_up[i-1]; | ||
| tmp_quotient = std::pow(2,i-1); |
There was a problem hiding this comment.
Maybe 1<<(i-1) is an easier and faster choice. pow in cmath is for float number
|
I will look at it @titansnow |
|
Anything new in this topic? Having a division operator would be great! |
|
@karlphillip Before going over any PR's I want to add an automated test suite to reduce regressions. Havent had any time lately; not sure if this is used in production anywhere so I havent rushed it. |
Adding the division operation.
Allows the division of two integer numbers of "Bigint" data type, returning quotient and remainder.