I can return a kj::Exception where a kj::Promise is expected, like so:
kj::Promise<void> someFunc() {
return KJ_EXCEPTION(FAILED, "some description");
}
But what if I end up having a kj::Promise<kj::Exception> in a place where I don't have a waitScope? For instance:
kj::Promise<void> someFunc(kj::Promise<void> somePromise) {
return somePromise.then([]() {
return KJ_EXCEPTION(FAILED, "some description");
});
}
There my compiler complains that there is no valid conversion between kj::Promise<kj::Exception> and kj::Promise<void>.
Is there a way around that?
I'm a little surprised that I didn't prohibit Promise<Exception> in the first place; using such a construction likely leads to a lot of problems.
You can avoid it in your code by explicitly declaring the return type of the lambda to be Promise<void>, i.e.:
return somePromise.then([]() -> kj::Promise<void> {
I am migrating my prototype from a listener to a visitor pattern. In the prototype, I have a grammar fragment like this:
thingList: thing+ ;
thing
: A aSpec # aRule
| B bSpec # bRule
;
Moving to a visitor pattern, I am not sure how I write visitThingList. Every visitor returns a specializes subclass of "Node", and I would love somehow when to be able to write something like this, say a "thingList" cares about the first thing in the list some how ...
visitThingList(cx: ThingListContext): ast.ThingList {
...
const firstThing = super.visit(cx.thing(0));
The problem with this is in typing. Each visit returns a specialized type which is a subclass of ast.Node. Because I am using super.visit, the return value will be the base class
of my node tree. However, I know because I am looking at the grammar
and because I wrote both vistARule and visitBRule that the result of the visit will be of type ast.Thing.
So we make visitThingList express it's expectation with cast ...
visitThingList(cx: ThingListContext): ast.ThingList {
const firstThing = super.visit(cx.thing(0));
if (!firstThing instanceof ast.Thing) {
throw "no matching visitor for thing";
}
// firstThing is now known to be of type ast.Thing
...
In much of my translator, type problems with ast Nodes are a compile time issue, I fix them in my editor. In this case, I am producing a more fragile walk, which will only reveal the fragility at runtime and then only with certain inputs.
I think I could change my grammar, to make it possible to encode the
type expectations of vistThingList() by creating a vistThing() entry point
thingList: thing+ ;
thing: aRule | bRule;
aRule: A aSpec;
bRule: B bSpec;
With vistThing() typed to match the expectation:
visitThing(cx: ThingContext): ast.Thing { }
visitThingList(cx: ThingListContext) {
const firstThing: ast.Thing = this.visitThing(cx.thing(0));
Now visitThingList can call this.visitThing() and the type enforcement of making sure all rules that a thing matches return ast.Thing belongs to visitThing(). If I do create a new rule for thing, the compiler will force me to change the return type of visitThing() and if I make it return something which is NOT a thing, visitThingList() will show type errors.
This also seems wrong though, because I don't feel like I should have to change my grammar in order to visit it.
I am new to ANTLR and wondering if there is a better pattern or approach to this.
When I was using the listener pattern, I wrote something like:
enterThing(cx: ThingContext) { }
enterARule(cx : ARuleContext) { }
enterBRule(cx : BRuleContext) { }
Not quite: for a labeled rule like thing, the listener will not contain enterThing(...) and exitThing(...) methods. Only the enter... and exit... methods for the labels aSpec and bSpec will be created.
How would I write the visitor walk without changing the grammar?
I don't understand why you need to change the grammar. When you keep the grammar like you mentioned:
thingList: thing+ ;
thing
: A aSpec # aRule
| B bSpec # bRule
;
then the following visitor could be used (again, there is no visitThing(...) method!):
public class TestVisitor extends TBaseVisitor<Object> {
#Override
public Object visitThingList(TParser.ThingListContext ctx) {
...
}
#Override
public Object visitARule(TParser.ARuleContext ctx) {
...
}
#Override
public Object visitBRule(TParser.BRuleContext ctx) {
...
}
#Override
public Object visitASpec(TParser.ASpecContext ctx) {
...
}
#Override
public Object visitBSpec(TParser.BSpecContext ctx) {
...
}
}
EDIT
I do not know how, as i iterate over that, to call the correct visitor for each element
You don't need to know. You can simply call the visitor's (super) visit(...) method and the correct method will be invoked:
class TestVisitor extends TBaseVisitor<Object> {
#Override
public Object visitThingList(TParser.ThingListContext ctx) {
for (TParser.ThingContext child : ctx.thing()) {
super.visit(child);
}
return null;
}
...
}
And you don't even need to implement all methods. The ones you don't implement, will have a default visitChildren(ctx) in them, causing (as the name suggests) all child nodes under them being traversed.
In your case, the following visitor will already cause the visitASpec and visitBSpec being invoked:
class TestVisitor extends TBaseVisitor<Object> {
#Override
public Object visitASpec(TParser.ASpecContext ctx) {
System.out.println("visitASpec");
return null;
}
#Override
public Object visitBSpec(TParser.BSpecContext ctx) {
System.out.println("visitBSpec");
return null;
}
}
You can test this (in Java) like this:
String source = "... your input here ...";
TLexer lexer = new TLexer(CharStreams.fromString(source));
TParser parser = new TParser(new CommonTokenStream(lexer));
TestVisitor visitor = new TestVisitor();
visitor.visit(parser.thingList());
I found this question on https://github.com/arialdomartini/Back-End-Developer-Interview-Questions#snippets
And I am curious about your opinion, I just can't find an decent solution of this refactor, and what pattern would apply in this very common case.
function()
{
HRESULT error = S_OK;
if(SUCCEEDED(Operation1()))
{
if(SUCCEEDED(Operation2()))
{
if(SUCCEEDED(Operation3()))
{
if(SUCCEEDED(Operation4()))
{
}
else
{
error = OPERATION4FAILED;
}
}
else
{
error = OPERATION3FAILED;
}
}
else
{
error = OPERATION2FAILED;
}
}
else
{
error = OPERATION1FAILED;
}
return error;
}
Do you have any idea of how to refactor this?
Actually, I feel there is way more space for refactoring than what suggested by Sergio Tulentsev.
The questions in the repo you linked are more about starting a conversation on code than closed-ended questions. So, I think it is worth to discuss the smells and design flaws of that code, to set up the refactoring goals.
Smells
I see these problems:
The code violates some of the SOLID principles. It surely violates the Open Closed Principle, as it is not possible to extend it without changing its code. E.g., adding a new operation would require adding a new if/else branch;
It also violate the Single Responsibility Principle. It just does too much. It performs error checks, it's responsible to execute all the 4 operations, it contains their implementations, it's responsible to check their results and to chain their execution in the right order;
It violates the Dependency Inversion Principle, because there are dependencies between high-level and low-level components;
It has a horrible Cyclomatic complexity
It exhibits high coupling and low cohesion, which is exactly the opposite of what is recommended;
It contains a lot of code duplication: the function Succeeded() is repeated in each branch; the structure of if/elses is replicated over and over; the assignment of error is duplicated.
It could have a pure functional nature, but it relies instead on state mutation, which makes reasoning about it not easy.
There's an empty if statement body, which might be confusing.
Refactoring
Let's see what could be done.
Here I'm using a C# implementation, but similar steps can be performed with whatever language.
I renamed some of the elements, as I believe honoring a naming convention is part of the refactoring.
internal class TestClass
{
HResult SomeFunction()
{
var error = HResult.Ok;
if(Succeeded(Operation1()))
{
if(Succeeded(Operation2()))
{
if(Succeeded(Operation3()))
{
if(Succeeded(Operation4()))
{
}
else
{
error = HResult.Operation4Failed;
}
}
else
{
error = HResult.Operation3Failed;
}
}
else
{
error = HResult.Operation2Failed;
}
}
else
{
error = HResult.Operation1Failed;
}
return error;
}
private string Operation1()
{
// some operations
return "operation1 result";
}
private string Operation2()
{
// some operations
return "operation2 result";
}
private string Operation3()
{
// some operations
return "operation3 result";
}
private string Operation4()
{
// some operations
return "operation4 result";
}
private bool Succeeded(string operationResult) =>
operationResult == "some condition";
}
internal enum HResult
{
Ok,
Operation1Failed,
Operation2Failed,
Operation3Failed,
Operation4Failed,
}
}
For the sake of simplicity, I supposed each operation returns a string, and that the success or failure is based on an equality check on the string, but of course it could be whatever. In the next steps, it would be nice if the code is independent from the result validation logic.
Step 1
It would be nice to start the refactoring with the support of some test harness.
public class TestCase
{
[Theory]
[InlineData("operation1 result", HResult.Operation1Failed)]
[InlineData("operation2 result", HResult.Operation2Failed)]
[InlineData("operation3 result", HResult.Operation3Failed)]
[InlineData("operation4 result", HResult.Operation4Failed)]
[InlineData("never", HResult.Ok)]
void acceptance_test(string failWhen, HResult expectedResult)
{
var sut = new SomeClass {FailWhen = failWhen};
var result = sut.SomeFunction();
result.Should().Be(expectedResult);
}
}
Our case is a trivial one, but being the quiz supposed to be a job interview question, I would not ignore it.
Step 2
The first refactoring could be getting rid of the mutable state: each if branch could just return the value, instead of mutating the variable error. Also, the name error is misleading, as it includes the success case. Let's just get rid of it:
HResult SomeFunction()
{
if(Succeeded(Operation1()))
{
if(Succeeded(Operation2()))
{
if(Succeeded(Operation3()))
{
if(Succeeded(Operation4()))
return HResult.Ok;
else
return HResult.Operation4Failed;
}
else
return HResult.Operation3Failed;
}
else
return HResult.Operation2Failed;
}
else
return HResult.Operation1Failed;
}
We got rid of the empty if body, making in the meanwhile the code slightly easier to reason about.
Step 3
If now we invert each if statement (the step suggested by Sergio)
internal HResult SomeFunction()
{
if (!Succeeded(Operation1()))
return HResult.Operation1Failed;
if (!Succeeded(Operation2()))
return HResult.Operation2Failed;
if (!Succeeded(Operation3()))
return HResult.Operation3Failed;
if (!Succeeded(Operation4()))
return HResult.Operation4Failed;
return HResult.Ok;
}
we make it apparent that the code performs a chain of executions: if an operation succeeds, the next operation is invoked; otherwise, the chain is interrupted, with an error. The GOF Chain of Responsibility Pattern comes to mind.
Step 4
We could move each operation to a separate class, and let our function receive a chain of operations to execute in a single shot. Each class would deal with its specific operation logic (honoring the Single Responsibility Principle).
internal HResult SomeFunction()
{
var operations = new List<IOperation>
{
new Operation1(),
new Operation2(),
new Operation3(),
new Operation4()
};
foreach (var operation in operations)
{
if (!_check.Succeeded(operation.DoJob()))
return operation.ErrorCode;
}
return HResult.Ok;
}
We got rid of the ifs altogether (but one).
Notice how:
The interface IOperation has been introduced, which is a preliminary move to decouple the function from the operations, complying the with the Dependency Inversion Principle;
The list of operations can easily be injected into the class, using the Dependency Injection.
The result validation logic has been moved to a separate class Check, injected into the main class (Dependency Inversion and Single Responsibility are satisfied).
internal class SimpleStringCheck : IResultCheck
{
private readonly string _failWhen;
public Check(string failWhen)
{
_failWhen = failWhen;
}
internal bool Succeeded(string operationResult) =>
operationResult != _failWhen;
}
We gained the ability to switch the check logic without modifying the main class (Open-Closed Principle).
Each operation has been moved to a separate class, like:
internal class Operation1 : IOperation {
public string DoJob()
{
return "operation1 result";
}
public HResult ErrorCode => HResult.Operation1Failed;
}
Each operation knows its own error code. The function itself became independent from it.
Step 5
There is something more to refactor on the code
foreach (var operation in operations)
{
if (!_check.Succeeded(operation.DoJob()))
return operation.ErrorCode;
}
return HResult.Ok;
}
First, it's not clear why the case return HResult.Ok; is handled as a special case: the chain could contain a terminating operation never failing and returning that value. This would allow us to get rid of that last if.
Second, our function still has 2 responsibility: to visit the chain, and to check the result.
An idea could be to encapsulate the operations into a real chain, so our function could reduce to something like:
return operations.ChainTogether(_check).Execute();
We have 2 options:
Each operation knows the next operation, so starting from operation1 we could execute the whole chain with a single call;
Operations are kept unaware of being part of a chain; a separate, encapsulating structure adds to operations the ability to be executed in sequence.
I'm going on with the latter, but that's absolutely debatable. I'm introducing a class modelling a ring in a chain, moving the code away from our class:
internal class OperationRing : IRing
{
private readonly Check _check;
private readonly IOperation _operation;
internal IRing Next { private get; set; }
public OperationRing(Check check, IOperation operation)
{
_check = check;
_operation = operation;
}
public HResult Execute()
{
var operationResult = _operation.DoJob();
if (_check.Succeeded(operationResult))
return Next.Execute();
return _operation.ErrorCode;
}
}
This class is responsible to execute an operation and to handle the execution to the next ring if it succeeded, or to interrupt the chain returning the right error code.
The chain will be terminated by a never-failing element:
internal class AlwaysSucceeds : IRing
{
public HResult Execute() => HResult.Ok;
}
Our original class reduces to:
internal class SomeClass
{
private readonly Check _check;
private readonly List<IOperation> _operations;
public SomeClass(Check check, List<IOperation> operations)
{
_check = check;
_operations = operations;
}
internal HResult SomeFunction()
{
return _operations.ChainTogether(_check).Execute();
}
}
In this case, ChainTogether() is a function implemented as an extension of List<IOperation>, as I don't believe that the chaining logic is responsibility of our class.
That's not the right answer
It's absolutely debatable that the responsibilities have been separated to the most appropriate classes. For example:
is chaining operations a task of our function? Or should it directly receive the chained structure?
why the use of an enumerable? As Robert Martin wrote in "Refactoring: Improving the Design of Existing Code": enums are code smells and should be refactored to polymorphic classes;
how much is too much? Is the resulting design too complex? Does the complexity of the whole application need this level of modularisation?
Therefore, I'm sure there are several other ways to refactor the original function. In a job interview, or in a pair programming session, I expect a lot of discussions and evaluations to occur.
You could use early returns here.
function() {
if(!SUCCEEDED(Operation1())) {
return OPERATION1FAILED;
}
if(!SUCCEEDED(Operation2())) {
return OPERATION2FAILED;
}
if(!SUCCEEDED(Operation3())) {
return OPERATION3FAILED;
}
if(!SUCCEEDED(Operation4())) {
return OPERATION4FAILED;
}
# everything succeeded, do your thing
return S_OK;
}
So I have a list and for each item in the list I have to "do something" based on each list element.
The list consists of codes and there are a total of 5 codes. The list may contain any or all codes.
So far I've used the forEach and i've written the if conditions inside it as below -
List<Category> categories = getCategories();
categories.stream().forEach(category -> {
if(category.equals(Category.A)) {
// do something
} else if(category.equals(Category.B)) {
// do something
} else if(category.equals(Category.C)) {
// do something
} else if(category.equals(Category.D)) {
// do something
} else if(category.equals(Category.E)) {
// do something
}
});
I'm looking at refactoring this. Can someone please look at how better this can be done?
The only thing I would improve is to use a switch-statement:
switch(category){
case Category.A:
// do Something
break;
}
As mentionend by luk2302 this will only work if Category is an enum.
You can add a doSomething method to the Category class and simply call it in the .forEach.
For example:
public class Category {
// any other methods/variable/constructors
public void doSomething() {
//do something
}
}
Then you can call it like this:
categories.stream().forEach(Category::doSomething);
If the // do something has not a common behaviour, you can move the if part inside the doSomething method.
At first, do not use multiline lambdas and create new method (with switch):
private void doSomethingBasedOnCategory(Category category) {
switch(category) {
case A: // do something
break;
case B: // do something
break;
case C: // do something
break;
case D: // do something
break;
case E: // do something
break;
}
}
Then use it in your lambda:
getCategories()
.stream()
.forEach(category -> doSomethingBasedOnCategory(category);
Another way is to create static map prefilled with keys (which will be Category.X) and values (which will be functions ready to use)
In JS, inner functions could be quite handy to simplify the code, something like:
function complexStuff() {
function step1() { ... }
function step2() { ... }
step1()
step2()
}
Is it possible to use something similar in Ruby, or are there different approaches?
I don't like private methods because private methods are available to the whole class, and in this case, I want to limit the scope of the inner function even more--to just one method.
A Ruby lambda is similar to an anonymous js function:
step1 = lambda {puts "I am a lambda!"}
step1.call "optional args", ...
Some shorthand:
f = -> {puts "Shorthand lambda"}
f.()
More info (including shorthand notation!!) here