Strange Compiler Anomaly: VC++/MinGW <resolved>

[Comreak]

--EDIT--
The problem turned out to be some "bad implemenation" for in_avail() in GCC.

Here's the code in question:

string input;

cin >> input; //lets say the input is "hello hello"
cout << "Your Text: " << input << endl;
cout << "Number of chars: " << cin.rdbuf()->in_avail() << endl;

When I compile this with Mingw, in_avail() returns 0. However, when I compile it with VC++7, it returns a proper count of the number of characters left in the buffer (in this case, 7). Any idea why this happens?

[twanvl]

in_avail returns the number of elements in the input buffer, as far as I know there is no guarantee that when you press a key that character goes directly to cin's buffer.
Why do you want to use in_avail anyway?

[Comreak]

Originally posted by twanvl
in_avail returns the number of elements in the input buffer, as far as I know there is no guarantee that when you press a key that character goes directly to cin's buffer.
Why do you want to use in_avail anyway?

If I wanted to clear the buffer, I'd use "cin.ignore(cin.rdbuf()->in_avail())".

I was just playing around and found out that something worked with one compiler and not with the other. Doesn't that strike you as a bit strange?

[Darkwraith]

I was just playing around and found out that something worked with one compiler and not with the other. Doesn't that strike you as a bit strange?

No. Compilers are strange creatures that vary. If you are not using the standard, then there is no gaurantee that what compiles on one, compiles on another. (I believe) pragmas are different from compiler to compiler. Sometimes, even how you include things is different (MSVC++6 vs. Borland 5.02)

[Comreak]

Well, I just checked the bug list over at http://gcc.gnu.org and apparently it wasn't a bug, just "poor implemenation". Hopefully, it'll get fixed in the next major release. Here's the report:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=7744