C# 4.0 optional out/ref arguments - c#-4.0

Does C# 4.0 allow optional out or ref arguments?

No.
A workaround is to overload with another method that doesn't have out / ref parameters, and which just calls your current method.
public bool SomeMethod(out string input)
{
...
}
// new overload
public bool SomeMethod()
{
string temp;
return SomeMethod(out temp);
}
If you have C# 7.0, you can simplify:
// new overload
public bool SomeMethod()
{
return SomeMethod(out _); // declare out as an inline discard variable
}
(Thanks #Oskar / #Reiner for pointing this out.)

As already mentioned, this is simply not allowed and I think it makes a very good sense.
However, to add some more details, here is a quote from the C# 4.0 Specification, section 21.1:
Formal parameters of constructors, methods, indexers and delegate types can be declared optional:
fixed-parameter:
attributesopt parameter-modifieropt type identifier default-argumentopt
default-argument:
= expression
A fixed-parameter with a default-argument is an optional parameter, whereas a fixed-parameter without a default-argument is a required parameter.
A required parameter cannot appear after an optional parameter in a formal-parameter-list.
A ref or out parameter cannot have a default-argument.

No, but another great alternative is having the method use a generic template class for optional parameters as follows:
public class OptionalOut<Type>
{
public Type Result { get; set; }
}
Then you can use it as follows:
public string foo(string value, OptionalOut<int> outResult = null)
{
// .. do something
if (outResult != null) {
outResult.Result = 100;
}
return value;
}
public void bar ()
{
string str = "bar";
string result;
OptionalOut<int> optional = new OptionalOut<int> ();
// example: call without the optional out parameter
result = foo (str);
Console.WriteLine ("Output was {0} with no optional value used", result);
// example: call it with optional parameter
result = foo (str, optional);
Console.WriteLine ("Output was {0} with optional value of {1}", result, optional.Result);
// example: call it with named optional parameter
foo (str, outResult: optional);
Console.WriteLine ("Output was {0} with optional value of {1}", result, optional.Result);
}

There actually is a way to do this that is allowed by C#. This gets back to C++, and rather violates the nice Object-Oriented structure of C#.
USE THIS METHOD WITH CAUTION!
Here's the way you declare and write your function with an optional parameter:
unsafe public void OptionalOutParameter(int* pOutParam = null)
{
int lInteger = 5;
// If the parameter is NULL, the caller doesn't care about this value.
if (pOutParam != null)
{
// If it isn't null, the caller has provided the address of an integer.
*pOutParam = lInteger; // Dereference the pointer and assign the return value.
}
}
Then call the function like this:
unsafe { OptionalOutParameter(); } // does nothing
int MyInteger = 0;
unsafe { OptionalOutParameter(&MyInteger); } // pass in the address of MyInteger.
In order to get this to compile, you will need to enable unsafe code in the project options. This is a really hacky solution that usually shouldn't be used, but if you for some strange, arcane, mysterious, management-inspired decision, REALLY need an optional out parameter in C#, then this will allow you to do just that.

ICYMI: Included on the new features for C# 7.0 enumerated here, "discards" is now allowed as out parameters in the form of a _, to let you ignore out parameters you don’t care about:
p.GetCoordinates(out var x, out _); // I only care about x
P.S. if you're also confused with the part "out var x", read the new feature about "Out Variables" on the link as well.

No, but you can use a delegate (e.g. Action) as an alternative.
Inspired in part by Robin R's answer when facing a situation where I thought I wanted an optional out parameter, I instead used an Action delegate. I've borrowed his example code to modify for use of Action<int> in order to show the differences and similarities:
public string foo(string value, Action<int> outResult = null)
{
// .. do something
outResult?.Invoke(100);
return value;
}
public void bar ()
{
string str = "bar";
string result;
int optional = 0;
// example: call without the optional out parameter
result = foo (str);
Console.WriteLine ("Output was {0} with no optional value used", result);
// example: call it with optional parameter
result = foo (str, x => optional = x);
Console.WriteLine ("Output was {0} with optional value of {1}", result, optional);
// example: call it with named optional parameter
foo (str, outResult: x => optional = x);
Console.WriteLine ("Output was {0} with optional value of {1}", result, optional);
}
This has the advantage that the optional variable appears in the source as a normal int (the compiler wraps it in a closure class, rather than us wrapping it explicitly in a user-defined class).
The variable needs explicit initialisation because the compiler cannot assume that the Action will be called before the function call exits.
It's not suitable for all use cases, but worked well for my real use case (a function that provides data for a unit test, and where a new unit test needed access to some internal state not present in the return value).

Use an overloaded method without the out parameter to call the one with the out parameter for C# 6.0 and lower. I'm not sure why a C# 7.0 for .NET Core is even the correct answer for this thread when it was specifically asked if C# 4.0 can have an optional out parameter. The answer is NO!

For simple types you can do this using unsafe code, though it's not idiomatic nor recommended. Like so:
// unsafe since remainder can point anywhere
// and we can do arbitrary pointer manipulation
public unsafe int Divide( int x, int y, int* remainder = null ) {
if( null != remainder ) *remainder = x % y;
return x / y;
}
That said, there's no theoretical reason C# couldn't eventually allow something like the above with safe code, such as this below:
// safe because remainder must point to a valid int or to nothing
// and we cannot do arbitrary pointer manipulation
public int Divide( int x, int y, out? int remainder = null ) {
if( null != remainder ) *remainder = x % y;
return x / y;
}
Things could get interesting though:
// remainder is an optional output parameter
// (to a nullable reference type)
public int Divide( int x, int y, out? object? remainder = null ) {
if( null != remainder ) *remainder = 0 != y ? x % y : null;
return x / y;
}

The direct question has been answered in other well-upvoted answers, but sometimes it pays to consider other approaches based on what you're trying to achieve.
If you're wanting an optional parameter to allow the caller to possibly request extra data from your method on which to base some decision, an alternative design is to move that decision logic into your method and allow the caller to optionally pass a value for that decision criteria in. For example, here is a method which determines the compass point of a vector, in which we might want to pass back the magnitude of the vector so that the caller can potentially decide if some minimum threshold should be reached before the compass-point judgement is far enough away from the origin and therefore unequivocally valid:
public enum Quadrant {
North,
East,
South,
West
}
// INVALID CODE WITH MADE-UP USAGE PATTERN OF "OPTIONAL" OUT PARAMETER
public Quadrant GetJoystickQuadrant([optional] out magnitude)
{
Vector2 pos = GetJoystickPositionXY();
float azimuth = Mathf.Atan2(pos.y, pos.x) * 180.0f / Mathf.PI;
Quadrant q;
if (azimuth > -45.0f && azimuth <= 45.0f) q = Quadrant.East;
else if (azimuth > 45.0f && azimuth <= 135.0f) q = Quadrant.North;
else if (azimuth > -135.0f && azimuth <= -45.0f) q = Quadrant.South;
else q = Quadrant.West;
if ([optonal.isPresent(magnitude)]) magnitude = pos.Length();
return q;
}
In this case we could move that "minimum magnitude" logic into the method and end-up with a much cleaner implementation, especially because calculating the magnitude involves a square-root so is computationally inefficient if all we want to do is a comparison of magnitudes, since we can do that with squared values:
public enum Quadrant {
None, // Too close to origin to judge.
North,
East,
South,
West
}
public Quadrant GetJoystickQuadrant(float minimumMagnitude = 0.33f)
{
Vector2 pos = GetJoystickPosition();
if (minimumMagnitude > 0.0f && pos.LengthSquared() < minimumMagnitude * minimumMagnitude)
{
return Quadrant.None;
}
float azimuth = Mathf.Atan2(pos.y, pos.x) * 180.0f / Mathf.PI;
if (azimuth > -45.0f && azimuth <= 45.0f) return Quadrant.East;
else if (azimuth > 45.0f && azimuth <= 135.0f) return Quadrant.North;
else if (azimuth > -135.0f && azimuth <= -45.0f) return Quadrant.South;
return Quadrant.West;
}
Of course, that might not always be viable. Since other answers mention C# 7.0, if instead what you're really doing is returning two values and allowing the caller to optionally ignore one, idiomatic C# would be to return a tuple of the two values, and use C# 7.0's Tuples with positional initializers and the _ "discard" parameter:
public (Quadrant, float) GetJoystickQuadrantAndMagnitude()
{
Vector2 pos = GetJoystickPositionXY();
float azimuth = Mathf.Atan2(pos.y, pos.x) * 180.0f / Mathf.PI;
Quadrant q;
if (azimuth > -45.0f && azimuth <= 45.0f) q = Quadrant.East;
else if (azimuth > 45.0f && azimuth <= 135.0f) q = Quadrant.North;
else if (azimuth > -135.0f && azimuth <= -45.0f) q = Quadrant.South;
else q = Quadrant.West;
return (q, pos.Length());
}
(Quadrant q, _) = GetJoystickQuadrantAndMagnitude();
if (q == Quadrant.South)
{
// Do something.
}

What about like this?
public bool OptionalOutParamMethod([Optional] ref string pOutParam)
{
return true;
}
You still have to pass a value to the parameter from C# but it is an optional ref param.

void foo(ref int? n)
{
return null;
}

Related

Groovy primitive double arithmetic

This yields 127
double middle = 255 / 2
While this yields 127.5
Double middle = 255 / 2
Meanwhile this yields 127.5 as well
double middle = (255 / 2) as double
I know that Groovy operates with BigDecimal per default, but to me this is a Huuge bug! How can this be?
This actually has nothing to do with BigDecimals, but rather with the type coercion from primitive integer to the primitive double. This problem is caused by the Groovy compiler and the (most probably) incorrect bytecode it produces. Take a look at the following bytecode representation of the first case. The following Groovy code:
void ex1() {
double x = 255 / 2
println x
}
gets compiled to a bytecode that can be represented as:
public void ex1() {
CallSite[] var1 = $getCallSiteArray();
double x = 0.0D;
if (BytecodeInterface8.isOrigInt() && BytecodeInterface8.isOrigD() && !__$stMC && !BytecodeInterface8.disabledStandardMetaClass()) {
int var5 = 255 / 2;
x = (double)var5;
} else {
Object var4 = var1[5].call(255, 2);
x = DefaultTypeTransformation.doubleUnbox(var4);
}
var1[6].callCurrent(this, x);
}
It shows that in this case, it is not possible to get 127.5 as a result, because the result of 255 / 2 expression is stored in the variable of type int. It feels like this is an example of inconsistent behavior because here is what the bytecode of the method that uses Double looks like:
public void ex2() {
CallSite[] var1 = $getCallSiteArray();
Double x = null;
if (BytecodeInterface8.isOrigInt() && !__$stMC && !BytecodeInterface8.disabledStandardMetaClass()) {
Object var4 = var1[8].call(255, 2);
x = (Double)ScriptBytecodeAdapter.castToType(var4, Double.class);
} else {
Object var3 = var1[7].call(255, 2);
x = (Double)ScriptBytecodeAdapter.castToType(var3, Double.class);
}
var1[9].callCurrent(this, x);
}
The main problem with this use case is that adding #TypeChecked does not prevent you from making this mistake - compilation passes and the incorrect result is returned. However, when we add #TypeChecked annotation to the method that uses Double the compilation error is thrown. Adding #CompileStatic solves the problem.
I've run some tests and I can confirm that this problem exists in the recent 2.5.6, as well as 3.0.0-alpha-4 versions. I've created a bug report in the Groovy JIRA project. Thanks for finding and reporting the problem!
UPDATE: Java does the same
It seems like this is not a Groovy bug - this is how Java does things as well. In Java, you can store a result of a division of two ints in the double variable, but you will get nothing else than an integer cast to the double. With {{Double}} type things are different in terms of the syntax but pretty similar in terms of the bytecode. With {{Double}} you need to explicitly cast at least one part of the equation to the {{double}} type, which results in the bytecode that casts both integers to the {{double}}. Consider the following example in Java:
final class IntDivEx {
static double div(int a, int b) {
return a / b;
}
static Double div2(int a, int b) {
return a / (double) b;
}
public static void main(String[] args) {
System.out.println(div(255,2));
System.out.println(div2(255,2));
}
}
When you run it you get:
127.0
127.5
Now, if you take a look at the bytecode it creates, you will see something like this:
//
// Source code recreated from a .class file by IntelliJ IDEA
// (powered by Fernflower decompiler)
//
final class IntDivEx {
IntDivEx() {
}
static double div(int a, int b) {
return (double)(a / b);
}
static Double div2(int a, int b) {
return (double)a / (double)b;
}
public static void main(String[] args) {
System.out.println(div(255, 2));
System.out.println(div2(255, 2));
}
}
The only difference (in terms of the syntax) between Groovy and Java is that Groovy allows you to implicitly cast an integer to Double, and that is why
Double x = 255 / 2
is the correct statement in Groovy, while Java, in this case, fails during the compilation with the following error:
Error:(10, 18) java: incompatible types: int cannot be converted to java.lang.Double
That is why in Java you need to use casting when you assign from integer to Double.

Creating a Templated Function to Fill a Vector with another depending on Size

Is there a base function in Rcpp that:
Fills entirely by a single value if size of a vector is 1.
Fills the other vector completely if same length.
Fills with an NA value if neither Vector are the same length nor a vector is of size 1.
I've written the above criteria as a function below using a NumericVector as an example. If there isn't a base function in Rcpp that performs said operations there should be a way to template the function so that given any type of vector (e.g. numeric, character and so on) the above logic would be able to be executed.
// [[Rcpp::export]]
NumericVector cppvectorize(NumericVector x,NumericVector y) {
NumericVector y_out(y.size());
if(x.size() == 1) {
for(int i = 0; i < y_out.size(); i++) {
y_out[i] = x[0];
}
} else if(x.size() == y_out.size()) {
for(int i = 0; i < y_out.size(); i++) {
y_out[i] = x[i];
}
} else {
for(int i = 0; i < y_out.size(); i++) {
y_out[i] = NA_REAL;
}
}
return y_out;
}
Unfortunately, the closest you will come to such a function is one of the rep variants that Rcpp supports. However, none of the variants match the desired output. Therefore, the only option is to really implement a templated version of your desired function.
To create the templated function, we will first create a routing function that handles the dispatch of SEXP objects. The rationale behind the routing function is SEXP objects are able to be retrieved from and surfaced into R using Rcpp Attributes whereas a templated version is not. As a result, we need to specify the SEXTYPE (used as RTYPE) dispatches that are possible. The TYPEOF() macro retrieves the coded number. Using a switch statement, we can dispatch this number into the appropriate cases.
After dispatching, we arrive at the templated function. The templated function makes use of the base Vector class of Rcpp to simplify the data flow. From here, the notable novelty will be the use of ::traits::get_na<RTYPE>() to dynamically retrieve the appropriate NA value and fill it.
With the plan in place, let's look at the code:
#include <Rcpp.h>
using namespace Rcpp;
// ---- Templated Function
template <int RTYPE>
Vector<RTYPE> vec_helper(const Vector<RTYPE>& x, const Vector<RTYPE>& y) {
Vector<RTYPE> y_out(y.size());
if(x.size() == 1){
y_out.fill(x[0]);
} else if (x.size() == y.size()) {
y_out = x;
} else {
y_out.fill(::traits::get_na<RTYPE>());
}
return y_out;
}
// ---- Dispatch function
// [[Rcpp::export]]
SEXP cppvectorize(SEXP x, SEXP y) {
switch (TYPEOF(x)) {
case INTSXP: return vec_helper<INTSXP>(x, y);
case REALSXP: return vec_helper<REALSXP>(x, y);
case STRSXP: return vec_helper<STRSXP>(x, y);
default: Rcpp::stop("SEXP Type Not Supported.");
}
// Need to return a value even though this will never be triggered
// to quiet the compiler.
return R_NilValue;
}
Sample Tests
Here we conduct a few sample tests on each of the supported data
# Case 1: x == 1
x = 1:5
y = 2
cppvectorize(x, y)
## [1] NA
# Case 2: x == y
x = letters[1:5]
y = letters[6:10]
cppvectorize(x, y)
## [1] "a" "b" "c" "d" "e"
# Case 3: x != y && x > 1
x = 1.5
y = 2.5:6.5
cppvectorize(x, y)
## [1] 1.5 1.5 1.5 1.5 1.5

Is there a name for a trampoline class?

I'm designing a programming language and one feature I'd like to add is a cross between a trampoline function and a class. That is, a class which takes in a literal akin to a generic class taking in a type. I'm stuck on a name for these because I haven't encountered them in a language before, is there something which already means this concept or something close? Using trampoline class is an option, but if there's something that more accurately describes this or is already in use in another language I'd prefer to go with it to cut down on the amount of jargon required in the documentation.
Pseudo-code follows to illustrate this principle in case it is not clear from the above:
class Point<const int n> {
private float[n] _value;
Point() {
for (int i = 0; i < n; i++) {
this._value[i] = 0f;
}
}
Point(Point<o> other) {
for (int i = 0; i < min(n, o); i++) {
this._value[i] = 0f;
}
}
public static float operator [index] (optional float value = null) {
if (value != null) { this._value[index] = value; }
return (this._value[index]);
}
public static Point<max(o, p)> operator + (Point<const int o> p1, Point<const int p> p2) {
Point<min(o, p)> small = (p1.n < p2.n ? p1 : p2);
Point<min(o, p)> large = (p1.n < p2.n ? p2 : p1);
Point<max(o, p)> ret = new Point<max(o, p)>(large);
for (int i = 0; i < min(o, p); i++) { ret[i] += small[i] }
return (ret);
}
}
The term you are looking for is dependent types. It means that a type cannot only have type parameters (like generics), but a type can also be parameterized with arbitrary values (the dependent type parameters). For example, you can define the signature of a function that takes a number n and returns an array of length n.
Sadly, dependent type checking in general is undecidable. This is, because you have to calculate the range of possible values of the dependent type parameters while the type checking itself is executed. To actually type check the program, you have to check whether two pieces of code produce the same range of possible values. This is known as extensional function equality and this is the part that is known to be undecidable in general.
Now, it might be true that dependent type checking becomes decidable if only compile-time constants are used as dependent type parameters. However, I am not sure about that.
In the comments below, we figured out that the part that seems to be the dependent type parameter should actually not be used for the type checking. Instead, it can be seen as an implicit parameter. It is similar to implicit parameter passing in the Scala programming language.

C++\Cli Parallel::For with thread local variable - Error: too many arguments

Trying to implement my first Parallel::For loop with a tread local variable to sum results of the loop. My code is based on an example listed in "Visual C++ 2010, by W. Saumweber, D. Louis (German). Ch. 33, P.804).
I get stuck in the implementation with syntax errors in the Parallel::For call. The errors are as follows, from left to right: a) expected a type specifier, b) too many arguments for generic class "System::Func", c) pointer to member is not valid for a managed class, d) no operator "&" matches these operands.
In line with the book, I create a collection with data List<DataStructure^> numbers, which is subject to a calculation performed in method computeSumScore which is called by the Parallel::For routine in method sumScore. All results are summed in method finalizeSumScore using a lock.
Below I paste the full code of the .cpp part of the class, to show what I have. The data collection "numbers" may look a bit messy, but that's due to organical growth of the program and me learning as I go along.
// constructor
DataCollection::DataCollection(Form1^ f1) // takes parameter of type Form1 to give acces to variables on Form1
{
this->f1 = f1;
}
// initialize data set for parallel processing
void DataCollection::initNumbers(int cIdx)
{
DataStructure^ number;
numbers = gcnew List<DataStructure^>();
for (int i = 0; i < f1->myGenome->nGenes; i++)
{
number = gcnew DataStructure();
number->concentrationTF = f1->myOrgan->cellPtr[cIdx]->concTFA[i];
number->stringA->AddRange(f1->myGenome->cStruct[i]->gString->GetRange(0, f1->myGenome->cChars));
number->stringB->AddRange(f1->myGenome->cStruct[i]->pString);
if (f1->myGenome->cStruct[i]->inhibitFunc)
number->sign = -1;
else
number->sign = 1;
numbers->Add(number);
}
}
// parallel-for summation of scores
double DataCollection::sumScore()
{
Parallel::For<double>(0, numbers->Count, gcnew Func<double>(this, &GenomeV2::DataCollection::initSumScore),
gcnew Func<int, ParallelLoopState^, double, double>(this, &GenomeV2::DataCollection::computeSumScore),
gcnew Action<double>(this, &GenomeV2::DataCollection::finalizeSumScore));
return summation;
}
// returns start value
double DataCollection::initSumScore()
{
return 0.0;
}
// perform sequence alignment calculation
double DataCollection::computeSumScore(int k, ParallelLoopState^ status, double tempVal)
{
int nwScore;
if (numbers[k]->concentrationTF > 0)
{
nwScore = NeedlemanWunsch::computeGlobalSequenceAlignment(numbers[k]->stringA, numbers[k]->stringB);
tempVal = Mapping::getLinIntMapValue(nwScore); // mapped value (0-1)
tempVal = (double) numbers[k]->sign * tempVal * numbers[k]->concentrationTF;
}
else
tempVal = 0.0;
return tempVal;
}
// locked addition
void DataCollection::finalizeSumScore(double tempVal)
{
Object^ myLock = gcnew Object();
try
{
Monitor::Enter(myLock);
summation += tempVal;
}
finally
{
Monitor::Exit(myLock);
}
}
Once this problem is solved I need to ensure that the functions called (computeGlobalSequenceAlignment and getLinIntMapvalue) are thread safe and the program doesn't get stalled on multiple treads accessing the same (static) variables. But this needs to work first.
Hope you can help me out.
Hans Passant answered my question in the comments (include full method name, add comma). Yet I cannot mark my question as answered, so this answer is to close the question.

How do I use groovy's AS keyword

This may be a duplicate but "as" is an INCREDABLY hard keyword to google, even S.O. ignores "as" as part of query.
So I'm wondering how to implement a class that supports "as" reflexively. For an example class:
class X {
private val
public X(def v) {
val=v
}
public asType(Class c) {
if (c == Integer.class)
return val as Integer
if(c == String.class)
return val as String
}
}
This allows something like:
new X(3) as String
to work, but doesn't help with:
3 as X
I probably have to attach/modify the "asType" on String and Integer somehow, but I feel any changes like this should be confined to the "X" class... Can the X class either implement a method like:
X fromObject(object)
or somehow modify the String/Integer class from within X. This seems tough since it won't execute any code in X until X is actually used... what if my first usage of X is "3 as X", will X get a chance to override Integer's asType before Groovy tries to call is?
As you say, it's not going to be easy to change the asType method for Integer to accept X as a new type of transformation (especially without destroying the existing functionality).
The best I can think of is to do:
Integer.metaClass.toX = { -> new X( delegate ) }
And then you can call:
3.toX()
I can't think how 3 as X could be done -- as you say, the other way; new X('3') as Integer is relatively easy.
Actually, you can do this:
// Get a handle on the old `asType` method for Integer
def oldAsType = Integer.metaClass.getMetaMethod( "asType", [Class] as Class[] )
// Then write our own
Integer.metaClass.asType = { Class c ->
if( c == X ) {
new X( delegate )
}
else {
// if it's not an X, call the original
oldAsType.invoke( delegate, c )
}
}
3 as X
This keeps the functionality out of the Integer type, and minimizes scope of the effect (which is good or bad depending on what you're looking for).
This category will apply asType from the Integer side.
class IntegerCategory {
static Object asType(Integer inty, Class c) {
if(c == X) return new X(inty)
else return inty.asType(c)
}
}
use (IntegerCategory) {
(3 as X) instanceof X
}

Resources