cs Oct 7

Goal : finish exception handling (ch5)

if time start chapter 9 , read all of ch 9 by next week, next quiz (after 5) is 9
int num = 1;
int denom = 0;

if (denom !=0)
    cout << num/denom;  (handling the error

exception handling : added to C++ baecause" separate detection of errors from handling of errors why ? --> the code that detects the error may not know the best way to handle it. eg.. display device being available. also to reuse code.

may be in a library of classes or functions written long before the application is written.

range_error is not thrown by default for out of range errors on subscript operators unless std_lib_facilities.h is used.



if you want the exception thrown use at function. v.at[2] as opposed to v[2] (which doesn't throw)

try {

}
catch (out_of_range&) {} // out_of_range errors
catch(...) {}  // all other errors

throw 0; // what is the type of the value 0 (int, so would be caught by ...

vector<int> v(3);

reference to v[10]  would not throw an out of range error but v.at[10] would throw an out of range error.

why does he use a struct sometimes and a class sometimes. He uses class instead of struct if there is a meaningful class invariant.  if there is no invariant then uses a struct instead of a class i.e. members can take any values or no constraints.

eg. a class for date (months could be only from Jan to Dec)

out_of_range : built in to standard library.

data members in range_error class  : index..
function in range_error class : constructor function.

class Bad_area{};  // do data member and member functions.. creates a default constructor.

 throw Bad_area() ; //creating a type.

catch(Bad_area)

postcondition would return the area. which should be positive so that may be checked...