Error handling in class method

Question

Suppose there is a class for working with the database. There are methods of this class that handle possible errors through if, writing the text of the error that occurred to the lastError class variable, sending a signal that an error has occurred and returning what the method should return in case of an error.

bool insertIntoSomeTable() { // нормальный ход выполнения программы if( ошибка подключения ) { lastError=ошибка подключения; emit errorHasOccured(lastError); return false; } // нормальный ход выполнения программы if( сервер не может принять данные) { lastError=сервер не может принять данные; emit errorHasOccured(lastError); return false; } // нормальный ход выполнения программы if( ошибка выполения запроса ) { lastError=ошибка выполения запроса; emit errorHasOccured(lastError); return false; } //и так далее return true }

As a result, these ifs are very cluttered with method code. Is it normal practice if I redefine my exception class for the database and rewrite the method like this:

  bool insertIntoSomeTable() { try { // нормальный ход выполнения программы if( ошибка подключения ) { throw DatabaseException("Ошибка подключения"); } // нормальный ход выполнения программы if( сервер не может принять данные) { throw DatabaseException("сервер не может принять данные"); } // нормальный ход выполнения программы if( ошибка выполения запроса ) { throw DatabaseException("ошибка выполения запроса"); } //и так далее return true } catch(DatabaseException& e) { lastError= e.what(); emit errorHasOccured(lastError); return false; } }

Also, of course, you can make this method void; in the method itself, do only throw, and catch errors somewhere higher. But I would not really like to do that. In general, is it ok to use the above option?

Majestio Majestio 2,917 one five 32 · Accepted Answer · 2017-01-28T10:47:05

Is it normal practice if I redefine my exception class for the database and rewrite the method like this

Will be normal.

Example of "wrong" ifs

 bool Func1() { if (err) return false; return true; } bool Func2() { if (err) return false; return true; } bool Func3() { if (err) return false; return true; } // ... int MainFunc() { if (!Func1()) return -1; if (!Func2()) return -2; if (!Func3()) return -3; }

In this example above, there is an explicit double check. First on the action, then on the return code. That's not necessary. It is for such a case, or rather - and including, that they invented exception handling.

It would be right

 void Func1() { if (err) throw myexception(1); } void Func2() { if (err) throw myexception(2); } void Func3() { if (err) throw myexception(3); } // ... int MainFunc() { try { Func1(); Func2(); Func3(); } catch(myexception &e) { // some } catch(...) { // some } }

But, if in your code "check-not-check" - write so as to eliminate verbosity. In general, we read Occam's Razor )

inherited its exception class from std :: exception (class DbException: std :: exception).
But, for some reason, exceptions from it are caught only if catching catch (DbException & e) explicitly, and if catching from a base catch (std :: exception & e), then it does not intercept it and everything ends with std :: terminate.
Create a separate question and give your code, otherwise you just have to guess

Error handling in class method

1 answer 1

More articles: