Recurring C++ and Qt anti-patterns

sierdzio

Small things from me:

lack of const in signal declarations. A signal will never modify an object so it can always be const. Qt will const_cast it away anyway, but it enables you to emit signals from const methods and (possibly) compiler to optimize a bit more
overuse of lambdas in slot connections even when a normal slot just makes more sense

Asperamanca

@sierdzio said in Recurring C++ and Qt anti-patterns:

Small things from me:

lack of const in signal declarations. A signal will never modify an object so it can always be const. Qt will const_cast it away anyway, but it enables you to emit signals from const methods and (possibly) compiler to optimize a bit more

overuse of lambdas in slot connections even when a normal slot just makes more sense

This brings me to a philosophical question: Do I want to be able to emit a signal from a const method, although the slot(s) attached to the signal may well modify data the originating const method could not itself modify?

sierdzio

@Asperamanca said in Recurring C++ and Qt anti-patterns:

This brings me to a philosophical question: Do I want to be able to emit a signal from a const method, although the slot(s) attached to the signal may well modify data the originating const method could not itself modify?

Yes, it's very debatable :D I did find a few occasions where it was useful (latest example: modifying behaviour of QTreeView without patching Qt - I have emitted a signal from const overloaded method and did my modifications there), but I agree it does not feel "right".

fcarney

Given the code:
modules.h

#ifndef MODULES_H
#define MODULES_H

#include <string>

void reg_module(int type, std::string name, int initedValue);

#endif // MODULES_H

modules.cpp

#include "modules.h"

using namespace std;

struct Modules
{
    Modules(): m_initedValue(0){}
    int m_type;
    string m_name;
    int m_initedValue;
} global_modules_struct[128];

void reg_module(int type, std::string name, int initedValue){
    global_modules_struct[type].m_type = type;
    global_modules_struct[type].m_name = name;
    global_modules_struct[type].m_initedValue = initedValue;
}

moduletype.h

#ifndef MODULETYPE_H
#define MODULETYPE_H

// nothing here

#endif // MODULETYPE_H

moduletype.cpp

#include "moduletype.h"
#include "modules.h"

struct SomeModule{
    SomeModule(){
        reg_module(10, "some type", 5);
    }
} SomeModuleInstance;

Ignore obvious indexing bounds checking issues for the array itself. Also ignore external array indexing possibly being out of bounds.

I just ran into a form of this problem in our code and it did not exhibit issues in Linux (that we know of) and did show issues in Windows. Linux used gcc and Windows used mingw. Same version of Qt 5.12.2 etc. Once identified it was really easy to see why this is a big issue.

Edit:
Technically global_modules_struct is not really global either. So ignore the misleading name.

Chris Kawa

I had to dig through this thing once, only the real code was like a hundred times longer and more convoluted.

// Library.h statically linked to and included in DLL and EXE
struct SomeType
{
   int typeId();
};
Q_DECLARE_METATYPE(SomeType);

// Library.cpp
int SomeType::typeId()
{
    return qMetaTypeId<SomeType>();
}

// main app
SomeType& var1 = getItFromDLL();
SomeType& var2 = getItFromEXE();

bool same = var1.typeId() == var2.typeId(); // nope

Pretty ugly thing to debug, especially since once in blue moon it actually works :/

Brunner2

This post is deleted!

Kent-Dorfman

@sierdzio said in Recurring C++ and Qt anti-patterns:

Yes, it's very debatable :D I did find a few occasions where it was useful (latest example: modifying behaviour of QTreeView without patching Qt - I have emitted a signal from const overloaded method and did my modifications there), but I agree it does not feel "right".

Actually, I have zero problem with this. The way my mind works it makes perfect sense, as the signal is a message to a receiving class (any class). It's not the sender method that modifies the object state. It is the message. My mind differentiates between the two.

Kent-Dorfman

@fcarney said in Recurring C++ and Qt anti-patterns:

Apparently the standard allows for it:
https://stackoverflow.com/questions/704466/why-doesnt-delete-set-the-pointer-to-null
The creator himself wonders why it isn't so. Its like C++ is this beautiful, amazing, and now, WILD animal roaming free in cyberspace... Yeah, maybe the analogy isn't all that great, but it does conjure up a cool picture.

Jumping back a few months on this one, but I think the decision to leave alone the pointer value upon an object delete is solid. If I understand the standard properly, the target of a delete can be an lvalue or and rvalue. So delete 0x34fc3d2200 should be a valid operation, right? How ya gonna change the value of an rvalue (in a traditional sense)?

Chris Kawa

Imagine clearing some sort of array:

for(type* ptr = some_array; something ; ++ptr)
{
     delete ptr;
}

Now imagine delete would zero that pointer. Do you see the problem? You would have to make another, temporary, pointer just so you zero the copy and your original doesn't get changed. In other words you're paying for what you don't use or even want. There's also problem of const pointers or pointers that you got from external APIs that do their own bookkeeping and might actually need that pointer value even after delete. It would create more problems than it solves.

fcarney

        int 🥩=1;
        int 🧀=1;
        int 🥬=1;
        int 🍞=1;
        int 🍅=1;
        int 🥪=🥩+🥬+🍅+🧀+🥩;
        cout << 🥪 << endl;

Fails to compile in C++17...

Kent-Dorfman

@fcarney said in Recurring C++ and Qt anti-patterns:

    int 🥩=1;
    int 🧀=1;
    int 🥬=1;
    int 🍞=1;
    int 🍅=1;
    int 🥪=🥩+🥬+🍅+🧀+🥩;
    cout << 🥪 << endl;

Fails to compile in C++17...

What's this "int" stuff? Doesn't the 17 standard deduce the type based on the rvalue? Not that I think that is necessarily a good thing though.

fcarney

@kent-dorfman said in Recurring C++ and Qt anti-patterns:

🍞

The real problem is this variable is unused.

Kent-Dorfman

@fcarney said in Recurring C++ and Qt anti-patterns:

The real problem is this variable is unused.

So in 17 unused variables are errors instead of warnings?

fcarney

@kent-dorfman said in Recurring C++ and Qt anti-patterns:

So in 17 unused variables are errors instead of warnings?

No, its just a bug in the code for a samich.

Kent-Dorfman

@fcarney said in Recurring C++ and Qt anti-patterns:

No, its just a bug in the code for a samich.

Samich... Are you a yinzer?

fcarney

@kent-dorfman said in Recurring C++ and Qt anti-patterns:

Are you a yinzer?

Had to look it up. Based on what I read, no. Not sure where I heard sandwich being called samich though. I am in western USA.

Chris Kawa

Being an a-hole as a recruiter:

What does o() mean?
What does o.o mean?
What does o->o mean?
What does o-->o mean?
What does o()--<=>--o() mean? Fun fact - crashes MSVC (yes, the compiler, not the compiled program)
What does [](){;o()++<=>++o();}() mean?

kshegunov

@chris-kawa said in Recurring C++ and Qt anti-patterns:

I had to dig through this thing once, only the real code was like a hundred times longer and more convoluted.
[Snip]
Pretty ugly thing to debug, especially since once in blue moon it actually works :/

Indeed. Although, this is windows specific. It works correctly on Linux as the symbol resolution happens at run time.

J.Hilk

@chris-kawa wow, some of those you don't encounter every day...

Chris Kawa

Yeah, it was a bit out of topic. Here's one anti-pattern I encounter something like 9/10 code reviews:

auto widget = new SomeWidget(some_widget);
auto layout = new SomeLayout(some_other_widget);
layout->addWidget(widget);

It's not a correctness bug. It's a subtle performance one. Compare this with:

auto widget = new SomeWidget();
auto layout = new SomeLayout();
layout->addWidget(widget);
some_other_widget->setLayout(layout);

If you don't see it - count how many times parents need to be changed and imagine there's not one but, say, 50 widgets and layouts.
For extra sweetness do the same when the parent widget is visible - how many times layouts need to be recalculated?