YATC (Yet Another Try Catch) (problem)

Hummm. I don't see any mistake. I'd say that either:
-foo() doesn't throw an exception
-you enter the catch(), but don't notice it because of some reason
-your compiler has a (major) bug. What compiler do you use?

Dani,

The syntax of the catch is in proper order. I extracted your code-snippet and worked it into a complete example with slight mocifications (see code below).

I suspect, as Gabriel said, that foo does not properly throw an exception, or that the type of exception thrown is incorrect and does not properly harmonize with your catch. In the modified code, foo throws a std::bad_cast and the catch catches everything. Look at the example. Perhaps it will help you track down the problem with your code.

Chris.

#include <string>
#include <iostream>
using namespace std;

class runtime_error
{
private:

const string str;

public:

runtime_error(const string& msg) : str(msg)
{
}
virtual ~runtime_error()
{
}
};

static const void foo()
{
cout << "foo" << endl;
std::bad_cast bc;
throw(bc);
}

class CDsp {
public:
class Error: public ::runtime_error
{
public:

Error ();
Error (const string& msg);

};
//rest of the class CDsp
};

/////////////////////////////////
//CDsp.cpp file:

CDsp::Error::Error (const string& msg)
:runtime_error (msg)
{ }

CDsp::Error::Error ()
:runtime_error ("ptu: error")
{}
///////////////////////////////////
//in main:
int main()
{
//code

try {
foo(); //throws exception type CDsp:Error
}
catch(.../*CDsp::Error &err*/)
{
cout << "catch" << endl;
// cout<<err.what()<<endl;
// cout<<"bla"<<endl; //catch(...)
}

//code

return 1;
}

Thanks for your replys.
I am tottaly sorry for not giving the context of foo() so here it is
void CDsp::foo() throw(Error)
{

cout<<"foo"<<endl;
throw(" foo err ");
}

In addition, should the probelm with the type be a problem when using catch(..)?
In debug mode, the throw starts, and the app aborts somewhere in libc while dealing with the string, never getting back to main and the catch().
Is my foo() throw falty?

I am using KDevelop 2.0 on Linux 2.4.18...

CPUWIzardd - I just saw your post now -
Could explain a bit more...
Thanks.

Thanks again
Dani.
PUWIzardd - I just saw your post now -
Could explain a bit more...
Thanks.

Paul,(in respect to your first post)

I have thought of that as well, but 2 things:
1. If this was the problem, a catch(...) with a cout<<"bla"<<endl; should of worked - and it doesn't.
2. what() is virtual (not pure) in calss exception which runtime_error is derived from.
The origenal what() does what I need - why should I overload it?

Can you explain - why the what() returns nothing?

Thanks

Dani. [

Originally posted by mikledet
Paul,(in respect to your first post)

I have thought of that as well, but 2 things:
1. If this was the problem, a catch(...) with a cout<<"bla"<<endl; should of worked - and it doesn't.
2. what() is virtual (not pure) in calss exception which runtime_error is derived from.
The origenal what() does what I need - why should I overload it?

Can you explain - why the what() returns nothing?

Thanks

Dani. [

Yeah, I made the post and did an edit later. For the record,
though, the cout << "bla" was commented out. <shrug> It's no
matter, though.

I came into the type of problem I discussed on another compiler
where there is no constructor handy.

--Paul