Home Page
Archive > Posts > 2009 > June

Picture of me at work
Everywhere I work is like a cave

One of my coworkers took this picture of me from his cell phone while at the office and I thought I’d share it. :-)

Jeffrey Riaboy at work
Virtual Functions in DLLs
And Shared Object Visibility

Since I am now working on making sure all my applications work in both Windows and Linux, I have been having to work a lot more with GCC recently (which I am basically totally switching too). It’s a bit trying having to learn an entirely different toolset after having used the same formula for many many years.

Linux shared objects are a bit different than Window’s DLLs in that all symbols are naturally exported instead of just the ones you specify, among other different behaviors like how the libraries are loaded and unloaded during runtime. The solution to the visibility problem is more recent as far as GCC goes, and is to use the visibility attribute. Symbol visibility in a shared library is very important for both symbol collision and library load time reasons. This means, however, that every functions/symbol needs to be marked as either not visible (for Linux), or exportable (for Windows).

I ran into a rather nasty problem however that I couldn’t find any information on when making a class exportable in a DLL. Basically, I had the following:

A.dll creation
//Determine whether we are exporting or importing functions from this DLL
#ifdef A_DLL
	#define DLLEXPORT __declspec(dllexport)
	#define DLLEXPORT __declspec(dllimport)

	int x;
	void a_foo();
	virtual void a_bar();
#define A_DLL
#include "A.h"
#include "stdio.h"
void a::a_foo() {printf("a_foo");}
void a::a_bar() {printf("a_bar");}
Compiling A.dll
g++ -c A.cpp -o A.o
g++ -shared -Wl,--output-def=libA.def -Wl,--out-implib=./libA.a -Wl,--dll A.o -o ./A.dll

B.dll creation
//Determine whether we are exporting or importing functions from this DLL
#ifdef B_DLL
	#define DLLEXPORT __declspec(dllexport)
	#define DLLEXPORT __declspec(dllimport)

	int x;
	void b_foo();
	virtual void b_bar();
#define B_DLL
#include "B.h"
#include "stdio.h"
void b::b_foo() {printf("b_foo");}
void b::b_bar() {printf("b_bar");}
Compiling B.dll
g++ -c B.cpp -o B.o
g++ -shared -Wl,--output-def=libB.def -Wl,--out-implib=./libB.a -Wl,--dll B.o -o ./B.dll

main.exe creation
#include "A.h"
#include "B.h"

int main()
	a a_var;

	b b_var;
	return 1;
Compiling main.exe
g++ -c main.cpp -o main.o
g++ -c main.o -o ./main.exe -L./ -lA -lB

The problem that occurred was when only 1 of the DLL files was included on Windows, everything worked fine, but when both were included, I could only access the virtual functions from one of the 2. I was getting an error about “required class export or error due to vtable not being found”.

After many hours of tinkering and fruitless research, I stumbled upon the solution by accident. It turns out I was using the incorrect syntax. The classes should have been defined as “class DLLEXPORT x {}” instead of “DLLEXPORT class x {}”. Why it worked when only 1 of the DLLs was present, but clobbered with multiple DLLs, I have no idea.

NULL Pointer for C++
Extending a language for what it’s lacking

I’ve recently been frustrated by the fact that NULL in C++ is actually evaluated as integral 0 instead of a pointer with the value of 0. This problem can be seen in the following example:

class String
	String(int i)  { /* ... */ } //Convert a number to a string
	String(char* i){ /* ... */ } //Copy a char* string directly into the class

String Foo(NULL); //This would give the string "Foo" the value "0" instead of a char* to (void*)0

The solution I came up with, which my good friend Will Erickson (aka Sarev0k) helped me revise, is as follows:

#undef NULL //If NULL is already defined, get rid of it
struct NULL_STRUCT { template <typename T> operator T*() { return (T*)0; } }; //NULL_STRUCT will return 0 to any pointer
static NULL_STRUCT NULL; //NULL is of type NULL_STRUCT and static (local to the current file)

After coming up with this way of doing it, I found out this concept is already a part of the new C++0x standard as nullptr, but since it is not really out yet, I still need a solution for the current C++ standard.

After getting this to work how I wanted it, I tested it out to make sure it is optimized correctly in compilers. When the compiler knows a value will be 0, it can apply lots of special assembly tricks.

Microsoft Visual C++ got it right by seeing that NULL was just 0 and applying appropriate optimizations, but GCC missed an optimization step and didn’t detect that it was 0 down the whole pipe. GCC, to my knowledge, however, isn’t exactly known for its optimization.

Example code:
BYTE* a=...; //Set a to an arbitrary value (best if brought in via an external method [i.e. stdin] so the compiler doesn’t make assumptions about the variable)
bool b=(a==NULL); //Set to b if a is 0 (NULL)
What MSVC6 outputs (and what it should be after optimization):
test eax,eax	//logical and a against itself to determine if it is 0 or not
sete al		//Set the lowest byte of eax to 1 if a is 0
What GCC gives
xor edx,edx	//Temporarily store 0 in edx for later comparison. This is a 0 trick, but 1 step higher than it could be used at.
cmp edx,eax	//Compare a against edx (0)
sete al		//Set the lowest byte of eax to 1 if a equals the value in edx

On a side note, it has been quite painful going from using assembly in Microsoft Visual C++ to GCC for 2 reasons:
  • I hate AT&T (as opposed to Intel) assembly syntax. It is rather clunky to use, and every program I’ve ever used is in Intel syntax (including all the Intel reference documentation). I tried turning on Intel syntax through a flag when compiling through GCC, but it broke GCC. :-\
  • Having to list which assembly registers are modified/used in the extended assembly syntax. This interface is also very clunky and, I have found, prone to bugs and problems.