Related
I am currently trying to figure out an alternative method for switch statements as the program I have the switch statements are getting really long and confusing. Therefore I thought it would be a good idea to use array of pointers to functions. I am using c++ and qt. But when I try and implement, I am getting the following error.
cannot convert 'CheckPl::comA' from type 'void (CheckPl::)()' to type 'void (*)()'
It would be much appreciated if someone would help me out with this or at least point me to correct direction.
[...] alternative method for switch statements as the program I have the switch statements are getting really long and confusing.
Extract each case block into a separate function; This way, the switch changes from a 10km long function to a dispatch function:
void dispatch_function()
{
switch(x)
{
case 1: do_case_1(); break;
...
case n: do_case_n(); break;
}
}
Therefore I thought it would be a good idea to use array of pointers to functions.
It's not a good idea (especially, not in the way you went about it - you are solving the xy problem). In C++, when you have a requirement for multiple functions that are called in similar conditions, you have the requirements for an abstract interface.
Your resulting client code should look like this:
std::vector<handlers> handlers; // filled with handler instances, one for each case
for(const auto& h: handlers) // replaces switch
if(h.fits_case(x)) // replaces case statement
{
h.do_case(x); // replaces case block
break;
}
It follows that your handler classes should inherit from a base class like this:
class handler_base
{
virtual bool fits_case(int x) = 0;
virtual void do_case(int x) = 0;
}
This is easy to understand (in both implementation and client code), it is modular, testable (you can test each case separately) and extensible (if you need a new case you only add the case and add it to the vector); It also doesn't use any pointers.
A pointer to a member function has to be stored in a variable of the appropriate type. A pointer to a member function is not compatible with a pointer to a function.
void (CheckPl::*mptr)() = &CheckPl::comA;
A pointer to a member function requires an instance to an object for invocation.
CheckPl c;
CheckPl *cp = &c;
(c.*mptr)();
(cp->*mptr)();
The hardest thing to remember about the above syntax is that the extra set of parentheses is required.
I extended the LLVM Kaleidoscope example to support Strings. I added a StringExprAST, which has the virtual Codegen method impl as follows:
Value *StringExprAST::Codegen() {
StringRef r(Val);
return ConstantDataArray::getString(getGlobalContext(), r, false);
}
I am trying to concatenate Strings and have a ConcatExprAST with its Codegen method. Upon trying to access the data in the ConstantDataArray, I need to cast the Value* back to a ConstantDataArray* in order to use the getAsString() method.
How can I do this?
Thanks for any help.
The proper way to cast any subtype of Value to another is via cast<>(), e.g.:
Value* v = ...
ConstantDataArray* result = cast<ConstantDataArray>(v);
However, keep in mind that in your example you don't return an object of type ConstantDataArray, but you return something of the return type of ConstantDataArray::getString(), which is not necessarily an instance of ConstantDataArray itself, all you know is that it's a Constant (for instance, it might be a ConstantAggregateZero, which is not a ConstantDataArray).
In any case, if you're not sure that v is indeed of a specific type, check it first with isa<>() (or use the dyn_cast<>() idiom), before performing the cast<>().
So I ran across this (IMHO) very nice idea of using a composite structure of a return value and an exception - Expected<T>. It overcomes many shortcomings of the traditional methods of error handling (exceptions, error codes).
See the Andrei Alexandrescu's talk (Systematic Error Handling in C++) and its slides.
The exceptions and error codes have basically the same usage scenarios with functions that return something and the ones that don't. Expected<T>, on the other hand, seems to be targeted only at functions that return values.
So, my questions are:
Have any of you tried Expected<T> in practice?
How would you apply this idiom to functions returning nothing (that is, void functions)?
Update:
I guess I should clarify my question. The Expected<void> specialization makes sense, but I'm more interested in how it would be used - the consistent usage idiom. The implementation itself is secondary (and easy).
For example, Alexandrescu gives this example (a bit edited):
string s = readline();
auto x = parseInt(s).get(); // throw on error
auto y = parseInt(s); // won’t throw
if (!y.valid()) {
// ...
}
This code is "clean" in a way that it just flows naturally. We need the value - we get it. However, with expected<void> one would have to capture the returned variable and perform some operation on it (like .throwIfError() or something), which is not as elegant. And obviously, .get() doesn't make sense with void.
So, what would your code look like if you had another function, say toUpper(s), which modifies the string in-place and has no return value?
Have any of you tried Expected; in practice?
It's quite natural, I used it even before I saw this talk.
How would you apply this idiom to functions returning nothing (that is, void functions)?
The form presented in the slides has some subtle implications:
The exception is bound to the value.
It's ok to handle the exception as you wish.
If the value ignored for some reasons, the exception is suppressed.
This does not hold if you have expected<void>, because since nobody is interested in the void value the exception is always ignored. I would force this as I would force reading from expected<T> in Alexandrescus class, with assertions and an explicit suppress member function. Rethrowing the exception from the destructor is not allowed for good reasons, so it has to be done with assertions.
template <typename T> struct expected;
#ifdef NDEBUG // no asserts
template <> class expected<void> {
std::exception_ptr spam;
public:
template <typename E>
expected(E const& e) : spam(std::make_exception_ptr(e)) {}
expected(expected&& o) : spam(std::move(o.spam)) {}
expected() : spam() {}
bool valid() const { return !spam; }
void get() const { if (!valid()) std::rethrow_exception(spam); }
void suppress() {}
};
#else // with asserts, check if return value is checked
// if all assertions do succeed, the other code is also correct
// note: do NOT write "assert(expected.valid());"
template <> class expected<void> {
std::exception_ptr spam;
mutable std::atomic_bool read; // threadsafe
public:
template <typename E>
expected(E const& e) : spam(std::make_exception_ptr(e)), read(false) {}
expected(expected&& o) : spam(std::move(o.spam)), read(o.read.load()) {}
expected() : spam(), read(false) {}
bool valid() const { read=true; return !spam; }
void get() const { if (!valid()) std::rethrow_exception(spam); }
void suppress() { read=true; }
~expected() { assert(read); }
};
#endif
expected<void> calculate(int i)
{
if (!i) return std::invalid_argument("i must be non-null");
return {};
}
int main()
{
calculate(0).suppress(); // suppressing must be explicit
if (!calculate(1).valid())
return 1;
calculate(5); // assert fails
}
Even though it might appear new for someone focused solely on C-ish languages, to those of us who had a taste of languages supporting sum-types, it's not.
For example, in Haskell you have:
data Maybe a = Nothing | Just a
data Either a b = Left a | Right b
Where the | reads or and the first element (Nothing, Just, Left, Right) is just a "tag". Essentially sum-types are just discriminating unions.
Here, you would have Expected<T> be something like: Either T Exception with a specialization for Expected<void> which is akin to Maybe Exception.
Like Matthieu M. said, this is something relatively new to C++, but nothing new for many functional languages.
I would like to add my 2 cents here: part of the difficulties and differences are can be found, in my opinion, in the "procedural vs. functional" approach. And I would like to use Scala (because I am familiar both with Scala and C++, and I feel it has a facility (Option) which is closer to Expected<T>) to illustrate this distinction.
In Scala you have Option[T], which is either Some(t) or None.
In particular, it is also possible to have Option[Unit], which is morally equivalent to Expected<void>.
In Scala, the usage pattern is very similar and built around 2 functions: isDefined() and get(). But it also have a "map()" function.
I like to think of "map" as the functional equivalent of "isDefined + get":
if (opt.isDefined)
opt.get.doSomething
becomes
val res = opt.map(t => t.doSomething)
"propagating" the option to the result
I think that here, in this functional style of using and composing options, lies the answer to your question:
So, what would your code look like if you had another function, say toUpper(s), which modifies the string in-place and has no return value?
Personally, I would NOT modify the string in place, or at least I will not return nothing. I see Expected<T> as a "functional" concept, that need a functional pattern to work well: toUpper(s) would need to either return a new string, or return itself after modification:
auto s = toUpper(s);
s.get(); ...
or, with a Scala-like map
val finalS = toUpper(s).map(upperS => upperS.someOtherManipulation)
if you don't want to follow a functional route, you can just use isDefined/valid and write your code in a more procedural way:
auto s = toUpper(s);
if (s.valid())
....
If you follow this route (maybe because you need to), there is a "void vs. unit" point to make: historically, void was not considered a type, but "no type" (void foo() was considered alike a Pascal procedure). Unit (as used in functional languages) is more seen as a type meaning "a computation". So returning a Option[Unit] does make more sense, being see as "a computation that optionally did something". And in Expected<void>, void assumes a similar meaning: a computation that, when it does work as intended (where there are no exceptional cases), just ends (returning nothing). At least, IMO!
So, using Expected or Option[Unit] could be seen as computations that maybe produced a result, or maybe not. Chaining them will prove it difficult:
auto c1 = doSomething(s); //do something on s, either succeed or fail
if (c1.valid()) {
auto c2 = doSomethingElse(s); //do something on s, either succeed or fail
if (c2.valid()) {
...
Not very clean.
Map in Scala makes it a little bit cleaner
doSomething(s) //do something on s, either succeed or fail
.map(_ => doSomethingElse(s) //do something on s, either succeed or fail
.map(_ => ...)
Which is better, but still far from ideal. Here, the Maybe monad clearly wins... but that's another story..
I've been pondering the same question since I've watched this video. And so far I didn't find any convincing argument for having Expected, for me it looks ridiculous and against clarity&cleanness. I have come up with the following so far:
Expected is good since it has either value or exceptions, we not forced to use try{}catch() for every function which is throwable. So use it for every throwing function which has return value
Every function that doesn't throw should be marked with noexcept. Every.
Every function that returns nothing and not marked as noexcept should be wrapped by try{}catch{}
If those statements hold then we have self-documented easy to use interfaces with only one drawback: we don't know what exceptions could be thrown without peeking into implementation details.
Expected impose some overheads to the code since if you have some exception in the guts of your class implementation(e.g. deep inside private methods) then you should catch it in your interface method and return Expected. While I think it is quite tolerable for the methods which have a notion for returning something I believe it brings mess and clutter to the methods which by design have no return value. Besides for me it is quite unnatural to return thing from something that is not supposed to return anything.
It should be handled with compiler diagnostics. Many compilers already emit warning diagnostics based on expected usages of certain standard library constructs. They should issue a warning for ignoring an expected<void>.
this is my first question after long time checking on this marvelous webpage.
Probably my question is a little silly but I want to know others opinion about this. What is better, to create several specific methods or, on the other hand, only one generic method? Here is an example...
unsigned char *Method1(CommandTypeEnum command, ParamsCommand1Struct *params)
{
if(params == NULL) return NULL;
// Construct a string (command) with those specific params (params->element1, ...)
return buffer; // buffer is a member of the class
}
unsigned char *Method2(CommandTypeEnum command, ParamsCommand2Struct *params)
{
...
}
unsigned char *Method3(CommandTypeEnum command, ParamsCommand3Struct *params)
{
...
}
unsigned char *Method4(CommandTypeEnum command, ParamsCommand4Struct *params)
{
...
}
or
unsigned char *Method(CommandTypeEnum command, void *params)
{
switch(command)
{
case CMD_1:
{
if(params == NULL) return NULL;
ParamsCommand1Struct *value = (ParamsCommand1Struct *) params;
// Construct a string (command) with those specific params (params->element1, ...)
return buffer;
}
break;
// ...
default:
break;
}
}
The main thing I do not really like of the latter option is this,
ParamsCommand1Struct *value = (ParamsCommand1Struct *) params;
because "params" could not be a pointer to "ParamsCommand1Struct" but a pointer to "ParamsCommand2Struct" or someone else.
I really appreciate your opinions!
General Answer
In Writing Solid Code, Steve Macguire's advice is to prefer distinct functions (methods) for specific situations. The reason is that you can assert conditions that are relevant to the specific case, and you can more easily debug because you have more context.
An interesting example is the standard C run-time's functions for dynamic memory allocation. Most of it is redundant, as realloc can actually do (almost) everything you need. If you have realloc, you don't need malloc or free. But when you have such a general function, used for several different types of operations, it's hard to add useful assertions and it's harder to write unit tests, and it's harder to see what's happening when debugging. Macquire takes it a step farther and suggests that, not only should realloc just do _re_allocation, but it should probably be two distinct functions: one for growing a block and one for shrinking a block.
While I generally agree with his logic, sometimes there are practical advantages to having one general purpose method (often when operations is highly data-driven). So I usually decide on a case by case basis, with a bias toward creating very specific methods rather than overly general purpose ones.
Specific Answer
In your case, I think you need to find a way to factor out the common code from the specifics. The switch is often a signal that you should be using a small class hierarchy with virtual functions.
If you like the single method approach, then it probably should be just a dispatcher to the more specific methods. In other words, each of those cases in the switch statement simply call the appropriate Method1, Method2, etc. If you want the user to see only the general purpose method, then you can make the specific implementations private methods.
Generally, it's better to offer separate functions, because they by their prototype names and arguments communicate directly and visibly to the user that which is available; this also leads to more straightforward documentation.
The one time I use a multi-purpose function is for something like a query() function, where a number of minor query functions, rather than leading to a proliferation of functions, are bundled into one, with a generic input and output void pointer.
In general, think about what you're trying to communicate to the API user by the API prototypes themselves; a clear sense of what the API can do. He doesn't need excessive minutae; he does need to know the core functions which are the entire point of having the API in the first place.
First off, you need to decide which language you are using. Tagging the question with both C and C++ here makes no sense. I am assuming C++.
If you can create a generic function then of course that is preferable (why would you prefer multiple, redundant functions?) The question is; can you? However, you seem to be unaware of templates. We need to see what you have omitted here to tell if you if templates are suitable however:
// Construct a string (command) with those specific params (params->element1, ...)
In the general case, assuming templates are appropriate, all of that turns into:
template <typename T>
unsigned char *Method(CommandTypeEnum command, T *params) {
// more here
}
On a side note, how is buffer declared? Are you returning a pointer to dynamically allocated memory? Prefer RAII type objects and avoid dynamically allocating memory like that if so.
If you are using C++ then I would avoid using void* as you don't really need to. There is nothing wrong with having multiple methods. Note that you don't actually have to rename the function in your first set of examples - you can just overload a function using different parameters so that there is a separate function signature for each type. Ultimately, this kind of question is very subjective and there are a number of ways of doing things. Looking at your functions of the first type, you would perhaps be well served by looking into the use of templated functions
You could create a struct. That's what I use to handle console commands.
typedef int (* pFunPrintf)(const char*,...);
typedef void (CommandClass::*pKeyFunc)(char *,pFunPrintf);
struct KeyCommand
{
const char * cmd;
unsigned char cmdLen;
pKeyFunc pfun;
const char * Note;
long ID;
};
#define CMD_FORMAT(a) a,(sizeof(a)-1)
static KeyCommand Commands[]=
{
{CMD_FORMAT("one"), &CommandClass::CommandOne, "String Parameter",0},
{CMD_FORMAT("two"), &CommandClass::CommandTwo, "String Parameter",1},
{CMD_FORMAT("three"), &CommandClass::CommandThree, "String Parameter",2},
{CMD_FORMAT("four"), &CommandClass::CommandFour, "String Parameter",3},
};
#define AllCommands sizeof(Commands)/sizeof(KeyCommand)
And the Parser function
void CommandClass::ParseCmd( char* Argcommand )
{
unsigned int x;
for ( x=0;x<AllCommands;x++)
{
if(!memcmp(Commands[x].cmd,Argcommand,Commands[x].cmdLen ))
{
(this->*Commands[x].pfun)(&Argcommand[Commands[x].cmdLen],&::printf);
break;
}
}
if(x==AllCommands)
{
// Unknown command
}
}
I use a thread safe printf pPrintf, so ignore it.
I don't really know what you want to do, but in C++ you probably should derive multiple classes from a Formatter Base class like this:
class Formatter
{
virtual void Format(unsigned char* buffer, Command command) const = 0;
};
class YourClass
{
public:
void Method(Command command, const Formatter& formatter)
{
formatter.Format(buffer, command);
}
private:
unsigned char* buffer_;
};
int main()
{
//
Params1Formatter formatter(/*...*/);
YourClass yourObject;
yourObject.Method(CommandA, formatter);
// ...
}
This removes the resposibility to handle all that params stuff from your class and makes it closed for changes. If there will be new commands or parameters during further development you don't have to modifiy (and eventually break) existing code but add new classes that implement the new stuff.
While not full answer this should guide you in correct direction: ONE FUNCTION ONE RESPONSIBILITY. Prefer the code where it is responsible for one thing only and does it well. The code whith huge switch statement (which is not bad by itself) where you need cast void * to some other type is a smell.
By the way I hope you do realise that according to standard you can only cast from void * to <type> * only when the original cast was exactly from <type> * to void *.
I've recently been doing a huge refactoring where I was changing a lot of my code to return booleans instead of an explicit return code. To aid this refactoring I decided to lean on the compiler where possible by getting it to tell me the places where my code needed to be changed. I did this by introducing the following class (see here for the lowdown on how this works):
///
/// Typesafe boolean class
///
class TypesafeBool
{
private:
bool m_bValue;
struct Bool_ {
int m_nValue;
};
typedef int Bool_::* bool_;
inline bool_ True() const { return &Bool_::m_nValue; }
inline bool_ False() const { return 0; }
public:
TypesafeBool( const bool bValue ) : m_bValue( bValue ){}
operator bool_() const { return m_bValue ? True() : False(); }
};
Now, instead of using a normal bool type as the return type, I used this class which meant that I couldn't compile something like this any more:
TypesafeBool SomeFunction();
long result = SomeFunction(); // error
Great: it has made the refactoring manageable on a huge codebase by letting the compiler do a lot of the hard work for me. So now I've finished my refactoring and I'd quite like to keep this class hanging around and carry on using it since it affords us an extra level of safety that the built-in bool type doesn't.
There is however one "problem" which is preventing me from doing this. At the moment we make heavy use of the ternary operator in our code, and the problem is that it is not compatible with this new class without explicit casts:
TypesafeBool result = ( 1 == 2 ? SomeFunction() : false ); // error: different types used
TypesafeBool result = ( 1 == 2 ? SomeFunction() : (TypesafeBool)false );
If I could "solve" this issue so that I could use my class in a seamless manner I would probably carry on using it throughout the codebase. Does anyone know of a solution to this issue? Or is it just impossible to do what I want?
In the context of the conditional operator, the type of the expression is the common type of the last two operands. The complete rules to determine this common type are a bit complex, but your case happens to be trivial: if one of the two possible return values is a class type, the other value must have the same class and the common type is obviously also that class.
That means that if one of the operands is a TypesafeBool, then the other must be as well.
Now the problem you're really trying to solve has been solved before. The trick is not providing a class; instead use a typedef. See for instance safe bool.
class CCastableToBool
{
public:
// ...
operator bool() const
{
//...
{
return true;
}
//...
return false;
}
private:
// ...
};
but beware, in C++ it is considered really dangerous to have a class that can be casted to bool. You are warned :-)
you can read this there, SafeBool
You should explicitely call TypesafeBool::True() in all your ternary tests.
TypesafeBool result = ( 1 == 2 ? SomeFunction().True() : false );
I don't know about a seamless manner, the ternary operator has some restrictions on its use...
However, why don't you define two constants ?
TypesafeBool const True = TypesafeBool(true);
TypesafeBool const False = TypesafeBool(false);
And then:
TypesafeBool result = ( 1 == 2 ? SomeFunction() : False );
Of course, it's a bit unorthodox since I play on the capitalization to avoid reusing a reserved word :)
Is it a possibility to make the constructor of TypesafeBool explicit? Of course, now the usage has to be
TypesafeBool result( 1 == 2 ? b : false );
Could you use an assignment operator that takes in a bool as the external argument, as well as one that takes a TypesafeBool? It might be something to try out...
Nice try, but if your code base is large, you are probably better off using a static checker such as PC-Lint to look for implicit bool<->int conversions instead.