This question already has answers here:
How to put assert into release builds in C/C++
(7 answers)
Closed 1 year ago.
Where can I find an ASSERT similar to the assert(...) macro from the standard C++ library (the one defined in <cassert>), but that works on Release mode too? Or how should I write one?
I like assert(...) because it automatically prints the line my source file where the assertion failed, as well as the assertion expression. I expect those features (if possible) in the Release mode ASSERT as well.
Basically assert is a macro that evaluates the expression and if it fails prints something and then aborts. It's not that hard to write something similar, something like.
#define ASSERT(x) do { if( !(x) ) { printfunc( #x ); abort(); } while(0)
Then you can modify that in order to suit your requirements. For example you might want not to abort in release mode. You could also adjust your printouts (to include just the information that you think is useful), in order to get file and line information you would use __FILE__ and __LINE__ macros (btw the #x in the define expands to a string literal containing the expression x).
In your release build, don't define NDEBUG, and assert will work.
You can undefine NDEBUG
#undef NDEBUG
near the assert statement
or you can define your own assert
#define assert(x) printf(...)
I've rolled out my own assertions that always fire in release mode too, based on this article:
This is the GIST of it, with me not providing the actual implementation (of AbstractAsserter, which is based on the article), but you get the idea:
#include <iostream>
struct AbstractAsserter
{
virtual void doAssert(const char* expr, const char* file, int line, const char* function) const
{
std::cout << "Asserting now at " << expr << ", " << file << ", " << line << ", " << function << std::endl;
}
};
struct Local
{
const char* function_;
const char* expr_;
const char* file_;
int line_;
Local( const char* f, const char* ex, const char* file, int line )
: function_( f ), expr_( ex ), file_( file ), line_( line )
{ }
Local operator << ( const AbstractAsserter& impl )
{
impl.doAssert( expr_, file_, line_, function_ );
return *this;
}
};
// More to be added as required...
#if defined( __GNUC__ )
# define WE_FUNCTION __PRETTY_FUNCTION__
#elif defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901)
# define WE_FUNCTION __func__
#else
# define WE_FUNCTION "null_func()"
#endif
#define WE_ASSERT( expr )\
if( expr );\
else Local( WE_FUNCTION, #expr, __FILE__, __LINE__ )\
<< AbstractAsserter();
int main() {
WE_ASSERT(!"bad BAD BOY!");
return 0;
}
With this structure, your implementation may also do some critical saving of state etc (if you deem your program state trustworthy...debatable).
Using the std assert in release mode (where, by the comments, you would like compiler optimizations enabled), isn't a problem per se, as is established by the other answers. Writing a custom one has been done, so there's not much problem there.
However, to quote the Dr. Dobb's article:
Do use assertions liberally throughout your code; they are watchful, reliable guards that protect you (your program) from insanity
If you are worried to have assertion + compiler optimizations, then these two thing are a bit add odds conceptually:
Asserts should and can be used liberally, with the intention of finding bugs early, and mostly so that it doesn't matter whether the assert code hurts performance.
If you want/need to run with optimizations, you may not like the performance hit the additional (now optimized, but still) assert code gives you.
So, in summary, I see the use of a relase mode assert, but make sure to keep it separate from the "normal" assert, where most people would assume it to be off in release mode.
Related
Let's say I have a project with many functions. For debugging purpopses, I want each one of them print out a diagnostic message (or a few) when called:
int f(int arg) {
cerr << "f() called with argument" << arg << endl;
int ret = 42;
cerr << "f() returned " << ret << endl;
return ret;
}
and so on. In one function, there may be as many as five to six messages printed out on cerr. I want to be able to disable them unless a debug flag (like NDEBUG) is set. One thing I could do is wrap each cerr statement with an if-statement.
#ifdef NDEBUG
const bool DEBUG_ON = true;
#else
const bool DEBUG_ON = false;
#endif
int f(int arg) {
if (DEBUG_ON) {
cerr << "f() called with argument" << arg << endl;
}
int ret = 42;
if (DEBUG_ON) {
cerr << "f() returned " << ret << endl;
}
return ret;
}
But it can get really tedious and, in the end, virtually unreadable with a couple dozens of such statements. My idea was to replace cerr with a custom object with an overloaded << operator which would send any arguments passed to it "into the void" like this:
class NullOutStream {
public:
NullOutStream operator<<(...) {
return *this;
}
};
NullOutStream debug_out;
#ifdef NDEBUG
#define debug_out cerr
#else
#define debug_out debug_out
#endif
Although it works perfectly, it doesn't seem very elegant. Is there a standard/nicer way to accomplish this?
There is no "standard" way to do this. There are many logging/assertion libraries that deal with similar issues, and there are different approaches to solve this. A few things to keep in mind:
It may or may not be desirable to keep the side effects of your log statement (as yours does). For example, if you did something like this:
debug_out << ++i;
This would increment i no matter whether you are in debug mode or not. This is probably desirable in this case. Conversely, if you do this:
debug_out << some_object.expensive_to_string_operation();
This would call expensive_to_string_operation() even when you are not in debug mode. This is probably not desirable in this case.
This would not happen if you used printf-style debug macro, e.g. something like this:
#ifdef NDEBUG
#define logf(...)
#else
#define logf(...) printf(__VA_ARGS__)
#endif
The reason being that when in non-debug mode, the arguments would be removed by the preprocessor.
You may want to do other things before or after each log statement, e.g. record a time stamp, flush a log file etc. It is possible to do this with streams by using destructor tricks, but it is more complicated to implement. It is much easier to do this using a function call, e.g. something like this:
#define log(msg) printf("%s: %s\n", timestamp(), msg)
You may want to record file name / line number with your logs. Again, this is easier to do with a function call than with a stream.
Streams may be better if you want custom formatting by object type - the printf interface doesn't lend itself to that very well.
I would recommend to have a look at some existing logging libraries to get an idea of different approaches. I suggest looking at Google's glog library because it has an interesting combination of using streams while retaining the ability to do 'per call' things (e.g. record time stamps, line numbers etc).
I am trying to write my own custom assert for my own project. This project will be written with c++11.
The assert must have the following qualities:
It must be kept as an expression and is assignable.
E.g. I should be able to write code like this int x = custom_assert(1/y);
It must be overloaded to accept an assert with a message and without one.
E.g int x = custom_assert(1/y, "Error divide by zero"); This code and the above are both compilable and acceptable.
It must have no side-effects in release mode
E.g. int x = custom_assert(1/y); will become int x = 1/y; in release mode.
And most importantly, it must break at the specific point where the assert was made. Which will make use of __debugbreak() as part of its evaluating expression.
The following is my attempt:
#include <string>
bool DoLog(std::string msg, std::string file, int line); //Prints to std::cerr and returns HALT macro
#if defined(_DEBUG) || defined(DEBUG)
#define HALT true
#define NT_ASSERT_BASE(x, msg) (!(x) && DoLog((msg), __FILE__, __LINE__) && (__debugbreak(),1))
#else
#define HALT false
#define NT_ASSERT_BASE(x,msg) (x)
#endif//debug/release defines
//--- Can't implement these until I can return the expression ---
//#define GET_MACRO(_1,_2,_3,NAME,...) NAME
//#define FOO(...) GET_MACRO(__VA_ARGS__, FOO3, FOO2)(__VA_ARGS__)
#define NT_ASSERT(expression, msg) NT_ASSERT_BASE(expression,msg)
As you can see my custom assert fails on 2 fronts, namely being kept as expression and assignable, and on overloading (Which I cannot implement until I figure out how to keep it as an expression.
All in all, I may be chasing stars and this macro may in fact be impossible to make. (Which I hope isn't the case)
Many thanks.
As far as I can tell, this can't be done in standard C++.
There is no way to get the __debugbreak() into the expanded code and at the same time pass the result of the expression unmodified, because you need the result twice: once for testing it, which will implicitly cast it to bool, and once to return it at the end.
There are two options:
Use gcc's and clang's ({}) construct with auto variable to hold the result. That will exclude MSC++, but I suppose you want that, because __debugbreak() is a MSC++ misfeature.
Give up on requiring the __debugbreak() on the call site, accept having to go one level up when it stops and make the thing as a template function.
A lambda expression will fit slightly better than a template function. It will make the break appear at the macro site, but it will still appear as a separate stack frame in the call stack. It also requires C++11 support (it was published over 5 years ago, but some platforms may not have it).
I don't think you should be mixing the validation with the assignment. From your example, it looks like you want to assign to an integer but an assertion, by nature, is a boolean expression. Further, your example is asserting on the wrong expression. It looks like you want to assert that y is not equal to zero (preventing division by zero), but you are asserting against something that will also be one or false or undefined.
If you are willing to be a bit flexible with your assignment requirements, then we can work around the problem of maintaining the expression and other useful info with some macro magic. Further, we can execute the __debugbreak() at the call site.
#include <iostream>
#include <string>
#include <type_traits>
template<class Fun>
inline bool DoLog(Fun f, std::string message, const char *expression, const char *filename, int line) {
static_assert(std::is_same<bool, decltype(f())>::value, "Predicate must return a bool.");
if (!(f())) {
std::cerr << filename << '#' << line << ": '" << expression << "' is false.";
if (!message.empty()) {
std::cerr << ' ' << message;
}
std::cerr << std::endl;
return false;
}
return true;
}
#if defined(_DEBUG) || defined(DEBUG)
#define HALT true
#define WITH_MESSAGE_(expr, x) [&](){return (expr);}, x, #expr
#define WITHOUT_MESSAGE_(expr) [&](){return (expr);}, std::string{}, #expr
#define PICK_ASSERTION_ARGS_(_1, _2, WHICH_, ...) WHICH_
#define CREATE_ASSERTION_ARGS_(...) PICK_ASSERTION_ARGS_(__VA_ARGS__, WITH_MESSAGE_, WITHOUT_MESSAGE_)(__VA_ARGS__)
#define NT_ASSERT(...) if (!DoLog(CREATE_ASSERTION_ARGS_(__VA_ARGS__), __FILE__, __LINE__)) __debugbreak()
#else
#define HALT false
#define NT_ASSERT(...)
#endif
int main() {
NT_ASSERT(true);
NT_ASSERT(false);
NT_ASSERT(1 == 1, "1 is 1");
NT_ASSERT(1 == 0, "1 is not 0");
return 0;
}
NOTE: The above snippet works on GCC using -std=c++11 (with a placeholder for the __debugbreak() statement). I'm making an assumption that VC++ would work also when it fully supports C++11.
I was programming a manchester decoding algorithm for arduino, and I often had to print debug stuff when trying to get things working, but printing to serial and string constants add a lot of overhead. I can't just leave it there in the final binary.
I usually just go through the code removing anything debug related lines.
I'm looking for a way to easily turn it on and off.
The only way I know is this
#if VERBOSE==1
Serial.println();
Serial.print(s);
Serial.print(" ");
Serial.print(t);
Serial.print(" preamble");
#endif
...
#if VERBOSE==1
Serial.println(" SYNC!\n");
#endif
and on top of the file I can just have
#define VERBOSE 0 // 1 to debug
I don't like how much clutter it adds to single liners. I was very tempted to do something very nasty like this. But yeah, evil.
Change every debug output to
verbose("debug message");
then use
#define verbose(x) Serial.print(x) //debug on
or
#define verbose(x) //debug off
There's a C++ feature that allows me to just do this instead of preprocessor?
At the risk of sounding silly: Yes, there is a C++ feature for this, it looks like this:
if (DEBUG)
{
// Your debugging stuff here…
}
If DEBUG is a compile-time constant (I think using a macro is reasonable but not required in this case), the compiler will almost certainly generate no code (not even a branch) for the debugging stuff if debug is false at compile-time.
In my code, I like having several debugging levels. Then I can write things like this:
if (DEBUG_LEVEL >= DEBUG_LEVEL_FINE)
{
// Your debugging stuff here…
}
Again, the compiler will optimize away the entire construct if the condition is false at compile-time.
You can even get more fancy by allowing a two-fold debugging level. A maximum level enabled at compile-time and the actual level used at run-time.
if (MAX_DEBUG >= DEBUG_LEVEL_FINE && Config.getDebugLevel() >= DEBUG_LEVEL_FINE)
{
// Your debugging stuff here…
}
You can #define MAX_DEBUG to the highest level you want to be able to select at run-time. In an all-performance build, you can #define MAX_DEBUG 0 which will make the condition always false and not generate any code at all. (Of course, you cannot select debugging at run-time in this case.)
However, if squeezing out the last instruction is not the most important issue and all your debugging code does is some logging, then the usual pattern lokks like this:
class Logger
{
public:
enum class LoggingLevel { ERROR, WARNING, INFO, … };
void logError(const std::string&) const;
void logWarning(const std::string&) const;
void logInfo(const std::string&) const;
// …
private:
LoggingLevel level_;
};
The various functions then compare the current logging level to the level indicated by the function name and if it is less, immediately return. Except in tight loops, this will probably be the most convenient solution.
And finally, we can combine both worlds by providing inline wrappers for the Logger class.
class Logger
{
public:
enum class LoggingLevel { ERROR, WARNING, INFO, … };
void
logError(const char *const msg) const
{
if (COMPILE_TIME_LOGGING_LEVEL >= LoggingLevel::ERROR)
this->log_(LoggingLevel::ERROR, msg);
}
void
logError(const std::string& msg) const
{
if (COMPILE_TIME_LOGGING_LEVEL >= LoggingLevel::ERROR)
this->log_(LoggingLevel::ERROR, msg.c_str());
}
// …
private:
LoggingLevel level_;
void
log_(LoggingLevel, const char *) const;
};
As long as evaluating the function arguments for your Logger::logError etc calls does not have visible side-effects, chances are good that the compiler will eliminate the call if the conditional in the inline function is false. This is why I have added the overloads that take a raw C-string to optimize the frequent case where the function is called with a string literal. Look at the assembly to be sure.
Personally I wouldn't have a a lot of #ifdef DEBUG scattered around my code:
#ifdef DEBUG
printf("something");
#endif
// some code ...
#ifdef DEBUG
printf("something else");
#endif
rather, I would wrap it in a function:
void DebugPrint(const char const *debugText) // ToDo: make it variadic [1]
{
#ifdef DEBUG
printf(debugText);
#endif
}
DebugPrint("something");
// some code ...
DebugPrint("something else");
If you don't define DEBUG then the macro preprocessor (not the compiler) won't expand that code.
The slight downside of my approach is that, although it makes your cod cleaner, it imposes an extra function call, even if DEBUG is not defined. It is possible that a smart linker will realize that the called function is empty and will remove the function calls, but I wouldn't bank on it.
References:
“Variadic function” in: Wikipedia, The Free Encyclopedia.
I also would suggest to use inline functions which become empty if a flag is set. Why when it is set? Because you usually want to debug always unless you compile a release build.
Because NDEBUG is already used you could use it too to avoid using multiple different flags. The definition of a debug level is also very useful.
One more thing to say: Be careful using functions which are altered by using macros! You could easily violate the One Definition Rule by translating some parts of your code with and some other without debugging disabled.
You might follow the convention of assert(3) and wrap debugging code with
#ifndef NDEBUG
DebugPrint("something");
#endif
See here (on StackOverflow, which would be a better place to ask) for a practical example.
In a more C++ like style, you could consider
ifdef NDEBUG
#define debugout(Out) do{} while(0)
#else
extern bool dodebug;
#define debugout(Out) do {if (dodebug) { \
std::cout << __FILE__ << ":" << __LINE__ \
<< " " << Out << std::endl; \
}} while(0)
#endif
then use debugout("here x=" << x) in your program. YMMV. (you'll set your dodebug flag either thru a gdb command or thru some program argument, perhaps parsed using getopt_long(3), at least on Linux).
PS. Remind that the do{...}while(0) is an old trick to make a robust statement like macro (suitable in every position where a plain statement is, e.g. as the then or else part of an if etc...).
You could also use templates utilizing the constexpr if feature in C++17. you don't have to worry about the preprocessor at all but your declaration and definition have to be in the same place when using templates.
Recently I have been reading Effective C++ Second Edition by Scott Meyers to improve on C++ best practices. One of his listed items encourages C++ programmers to avoid pre-processor macros and 'prefer the compiler'. He went as far as saying there are almost no reasons for macro in C++ aside from #include and #ifdef/#ifndef.
I agree with his reasoning, as you can accomplish the following macro
#define min(a,b) ((a) < (b) ? (a) : (b))
with the following C++ language features
template<class T>
inline const T & min(const T & a, const T & b) {
return a < b ? a : b;
}
where inline gives the compiler the option to remove the function call and insert inline code and template which can handle multiple data types who have an overloaded or built in > operator.
EDIT-- This template declaration will not completely match the stated macro if the data type of a and b differ. See Pete's comment for an example.
However, I am curious to know if using macros for debug logging is a valid use in C++. If the method I present below is not good practice, would someone be kind to suggest an alternative way?
I have been coding in Objective-C for the last year and one of my favorite 2D engines (cocos2d) utilized a macro to create logging statements. The macro is as follows:
/*
* if COCOS2D_DEBUG is not defined, or if it is 0 then
* all CCLOGXXX macros will be disabled
*
* if COCOS2D_DEBUG==1 then:
* CCLOG() will be enabled
* CCLOGERROR() will be enabled
* CCLOGINFO() will be disabled
*
* if COCOS2D_DEBUG==2 or higher then:
* CCLOG() will be enabled
* CCLOGERROR() will be enabled
* CCLOGINFO() will be enabled
*/
#define __CCLOGWITHFUNCTION(s, ...) \
NSLog(#"%s : %#",__FUNCTION__,[NSString stringWithFormat:(s), ##__VA_ARGS__])
#define __CCLOG(s, ...) \
NSLog(#"%#",[NSString stringWithFormat:(s), ##__VA_ARGS__])
#if !defined(COCOS2D_DEBUG) || COCOS2D_DEBUG == 0
#define CCLOG(...) do {} while (0)
#define CCLOGWARN(...) do {} while (0)
#define CCLOGINFO(...) do {} while (0)
#elif COCOS2D_DEBUG == 1
#define CCLOG(...) __CCLOG(__VA_ARGS__)
#define CCLOGWARN(...) __CCLOGWITHFUNCTION(__VA_ARGS__)
#define CCLOGINFO(...) do {} while (0)
#elif COCOS2D_DEBUG > 1
#define CCLOG(...) __CCLOG(__VA_ARGS__)
#define CCLOGWARN(...) __CCLOGWITHFUNCTION(__VA_ARGS__)
#define CCLOGINFO(...) __CCLOG(__VA_ARGS__)
#endif // COCOS2D_DEBUG
This macro provides for incredible utility which I will want to incorporate in my C++ programs. Writing a useful log statement is as simple as
CCLOG(#"Error in x due to y");
What is even better, is that if the COCOS2D_DEBUG is set to 0, then these statements never see the light of day. There is no overhead for checking a conditional statement to see if logging statements should be used. This is convenient when making the transition from development to production. How could one recreate this same effect in C++?
So does this type of macro belong in a C++ program? Is there a better, more C++ way of doing this?
First, Scott's statement was made at a time when macros were
considerably overused, for historical reasons. While it is
generally true, there are a few cases where macros make sense.
Of of these is logging, because only a macro can automatically
insert __FILE__ and __LINE__. Also, only a macro can
resolve to absolutely nothing (although based on my experience,
this isn't a big deal).
Macros such as you show aren't very frequent in C++. There are
two usual variants for logging:
#define LOG( message ) ... << message ...
which allows messages in the form " x = " << x, and can be
completely suppressed by redefining the macro, and
#define LOG() logFile( __FILE__, __LINE__ )
where logFile returns a wrapper for an std::ostream, which
defines operator<<, and permits such things as:
LOG() << "x = " << x;
Done this way, all of the expressions to the right of LOG()
will always be evaluated, but done correctly, no formatting will
be done unless the log is active.
There are "right" things to use macros for and there are bad uses of macros. Using macros where functions work is a bad idea. Using macros where functions DON'T do the same thing is perfectly good in my book.
I quite often use constructs like this:
#defien my_assert(x) do { if (!x) assert_failed(x, #x, __FILE__, __LINE__); } while(0)
template<typename T>
void assert_failed(T x, const char *x_str, const char *file, int line)
{
std::cerr << "Assertion failed: " << x_str << "(" << x << ") at " << file << ":" << line << std::endl;
std::terminate();
}
Another trick using the stringizing "operator" is something like this:
enum E
{
a,
b,
c,
d
};
struct enum_string
{
E v;
const char *str;
};
#define TO_STR(x) { x, #x }
enum_string enum_to_str[] =
{
TO_STR(a),
TO_STR(b),
TO_STR(c),
TO_STR(d),
};
Saves quite a bit of repeating stuff...
So, yes, it's useful in some cases.
After reading another question about the use of macros, I wondered: What are they good for?
One thing I don't see replaced by any other language construct very soon is in diminishing the number of related words you need to type in the following:
void log_type( const bool value ) { std::cout << "bool: " << value; }
void log_type( const int value ) { std::cout << "int: " << value; }
...
void log_type( const char value ) { std::cout << "char: " << value; }
void log_type( const double value ) { std::cout << "int: " << value; }
void log_type( const float value ) { std::cout << "float: " << value; }
as opposed to
#define LOGFN( T ) void log_type( const T value ) { std::cout << #T ## ": " << value; }
LOGFN( int )
LOGFN( bool )
...
LOGFN( char )
LOGFN( double )
LOGFN( float )
Any other 'irreplaceables'?
EDIT:
trying to summarize the reasons-why encountered in the answers; since that's what I was interested in. Mainly because I have a feeling that most of them are due to us still programming in raw text files in, still, poorly supporting environments.
flexibility of code-to-be compiled (e.g. #ifdef DEBUG, platform issues) (SadSido, Catalin, Goz)
debug purposes (e.g. __LINE__, __TIME__); I also put 'stringifying' under this reason (SadSido, Jla3ep, Jason S)
replacing e.g. PHP's require vs. include feature (#pragma once) (SadSido, Catalin)
readability enhancement by replacing complicated code (e.g. MESSAGEMAP, BOOST_FOREACH) (SadSido, fnieto)
DRY principle (Jason S)
an inline replacement (Matthias Wandel, Robert S. Barnes)
stringifying (Jason S)
compile different code under different conditions ( #ifdef __DEBUG__ );
guards to include each header once for each translation unit ( #pragma once );
__FILE__ and __LINE__ - replaced by the current file name and current line;
structuring the code to make it more readable (ex: BEGIN_MESSAGE_MAP() );
See interesting macro discussion at gotw here:
http://www.gotw.ca/gotw/032.htm
http://www.gotw.ca/gotw/077.htm
Most useful - header file guarding:
#ifndef MY_HEADER_GUARD
#define MY_HEADER_GUARD
// Header file content.
#endif
Later add [Windows only]
Exporting classes to DLL:
#ifdef EXPORT_MY_LIB
#define MY_API __declspec( dllexport)
#else
#define MY_API __declspec( dllimport)
#endif
Sample class:
class MY_API MyClass { ... };
platform specific sections.
ie
#ifdef WINDOWS
#include "WindowsImplementation.h"
#elif defined( LINUX )
#include "LinuxImplementation.h"
#else
#error Platform undefined.
#endif
I've posted this before, but of course cannot now find it. If you want to access the __FILE__ and __LINE__ macros, then another macro is by far the most convenient way to go - for example:
#define ATHROW(msg) \
{ \
std::ostringstream os; \
os << msg; \
throw ALib::Exception( os.str(), __LINE__, __FILE__ ); \
}
For doing cool magic tricks like in BOOST_FOREACH, injecting variables into an ambit.
BOOST_FOREACH( char c, "Hello, world!" )
{
... use char variable c here ...
} // c's scope ends here
// if there's an outer c defined, its scope resumes here
For don't-repeat-yourself (DRY) reasons. Things that involve repeated constructs at compile-time which cannot be abstracted away in other methods (templates or what have you). If you are finding you're repeating the same code constructs 20 times, that's a potential source of human error -- which hopefully can be abstracted away using templates but sometimes not. It's always a balance between the advantages of seeing raw code that can be type-checked and reviewed clearly, vs. the advantages of using macros for arbitrary substitution patterns (that generally can't be checked by automatic programming tools).
Stringifying and concatenation (the # and ## preprocessor patterns) can't be performed by templates.
Of course, at some point you may be better off using a tool (whether custom or off-the-shelf) for automatic code generation.
Modern languages take the philosophy that needing a proeprocessor ins a sign of a missing language feature, so they define all kinds of language features that the preprocessor took care of very simply back in the old K&R style C.
Your code example above could be simplified via an inline function, for example.
Personally, the most indispensable aspect of a preprocessor is to make it clear that some things are done compile time right in the source code. The java approach of eliminating dead code paths at compile time is just not as obvious when reading the code.
One of their uses is basically as a poor mans ( inlined ) template function.
For example:
#define MIN(X,Y) ((X) < (Y) ? : (X) : (Y))
This allows you to generate a custom MIN function for any types supporting these operators which is effectively inline at the point of useage. Of course there is no type checking and it's easy to end up with weird syntax errors or incorrect behavior if you don't get the parens just right.