#include <iostream> using namespace std; Int main() { cout<<"Hello World"; return 0; }

I'm a newbie to c++ and I'm faced with lots of error messages. I don't know what to do;Please guys your assistance on this is highly needed🙏🏽

I'm a newbie to c++ and I'm faced with lots of error messages. I don't know what to do;Please guys your assistance on this is highly needed🙏🏽

Bro in starting u will face error only Don't think too much it's normal Try to find out why those errors are coming

#include <iostream> using namespace std; Int main() { cout<<"Hello World"; return 0; }

#include <iostream> using namespace std; int main() { cout << "Hello World"; return 0; }

U have written Int but u have to write it int because it is a keyword

/start@MissRose_bot

No, a constexpr variable must be initialized with a constant expression So, it's always initialized at compile-time and can be used in other constant expressions So, it's always evaluated at compile-time

I believe constexpr conveys always only a possibility to be evaluated at compile-time, unless you use static constexpr. And yes, consteval is not for variables or static data members.

Can't template because I can theoretically store tens or hundreds of permutations of types, wouldn't want to make all them as template arguments of the class

Nice point.. but you have to use it judiciously.. If you're thinking it will explode the no of functions.. it won't if you design it properly.. using type erasure and wrap stuff.. basically any works like this.. so suppose you should use any over raw pointers..

And the other thing you should almost never prefer void* or std::any to templates

Isn't std::any a template based wrapper ? Why not use that? Edit: ok coz value copy ..

Isn't std::any a template based wrapper ? Why not use that? Edit: ok coz value copy ..

i think std::any is not a template but a type erasured type?

i think std::any is not a template but a type erasured type?

Yeah.. that in itself would've been alright for most use cases... But it allocates memory and copies value.. but prob someone out there wrote a reference copy version

I believe constexpr conveys always only a possibility to be evaluated at compile-time, unless you use static constexpr. And yes, consteval is not for variables or static data members.

Constexpr variables are never executed at runtime, even if the standard says they can

Hi everyone. Currently i am digging into DPDK source code rte_ring. This ring allows to enqueu/dequeue lock-free for multiple producers/consumers. The question is, why is there prod_tail atomic? What does it prevent us from?

This could help (look at the part titled Disruptor) https://archivist.nekoit.xyz/fast-logging/

This could help (look at the part titled Disruptor) https://archivist.nekoit.xyz/fast-logging/

Thank you, but i already know the answer. Producer tail used in calculation of max number of elements that are ready to be dequeued

I believe constexpr conveys always only a possibility to be evaluated at compile-time, unless you use static constexpr. And yes, consteval is not for variables or static data members.

No sane compiler would push errors to be detected at runtime, and if a compiler was dumb enough evaluate constexpr variables at runtime, it would also require template evaluation at runtime, as constexpr variables are absolutely usable as template arguments. It would also make constexpr variables unusable in static assertions

Thank you, but i already know the answer. Producer tail used in calculation of max number of elements that are ready to be dequeued

Yes, basically, you use several atomics to delimitate different regions that are interacted upon differently

Yes, basically, you use several atomics to delimitate different regions that are interacted upon differently

Are you familar with dpdk rte_ring primitive? Just i have got some other question about that

Are you familar with dpdk rte_ring primitive? Just i have got some other question about that

I have not used it in a few years but go on

Constexpr variables are never executed at runtime, even if the standard says they can

What's the resource for that claim?

What's the resource for that claim?

I've already explain you why

https://t.me/programminginc/500668 https://t.me/programminginc/500669 This is the standard ☝️

I've already explain you why

Because of constant literals?

https://t.me/programminginc/500670 This is basic knowledge about compilers

If the reason is that, so something like int a {5}; will also be evaluated at compile-time since 5 is a constant literal!

I'm shocked you neither agree with standards nor cppref!

If the reason is that, so something like int a {5}; will also be evaluated at compile-time since 5 is a constant literal!

If it's not changed during the execution, it also will be propagated Yet a itself is not a constant expression

I'm shocked you neither agree with standards nor cppref!

Lol, I just pointed to the standard

But the standard is not the "truth" per se

You live in a real world while the standard is not

I'm shocked you neither agree with standards nor cppref!

Can you please give at least one example when a constexpr variable is not evaluated at compile-time? Real world example

Also, appealing to authority is a form of logic error, so "Turner said so" is not an argument

Can you please give at least one example when a constexpr variable is not evaluated at compile-time? Real world example

There's a subtle difference in between. I didn't say a constexpr variable cannot be evaluated at compile time; I said, or better, by standard, it can be and most of the time it is in practice. To put in a nutshell, it's an opportunity not certainty.

There's a subtle difference in between. I didn't say a constexpr variable cannot be evaluated at compile time; I said, or better, by standard, it can be and most of the time it is in practice. To put in a nutshell, it's an opportunity not certainty.

That's not what I asked

I said that it is ALWAYS evaluated at compile-time You said that's not true So I ask you for a real-world of a constexpr variable evaluated at runtime

C++ is defined not only by the standard but by the real world too It's time to "turn on" your head and think why the fuck a thing that ALWAYS initialized at compile-time (that's the standard btw), we'll be sometimes evaluated at runtime.

C++ is defined not only by the standard but by the real world too It's time to "turn on" your head and think why the fuck a thing that ALWAYS initialized at compile-time (that's the standard btw), we'll be sometimes evaluated at runtime.

You can be happy with your head so-called turned on while your claims don't apply with standards. I tend to rely upon standard. Enough

You can be happy with your head so-called turned on while your claims don't apply with standards. I tend to rely upon standard. Enough

Do you understand that there's no implementation of C++ that actually aligns with the standard?

It's so foolish "to rely on the standard" because C++ is defined by both the standard and the real world, as I already said

Anyways, living in an imaginary world must be fun

Nice point.. but you have to use it judiciously.. If you're thinking it will explode the no of functions.. it won't if you design it properly.. using type erasure and wrap stuff.. basically any works like this.. so suppose you should use any over raw pointers..

Well "design it properly" also means not to overcomplicate. I see trying to move everything to compile time is overcomplication here. Type erasure yes, but that's what I'm basically implementing with void*. Note that void* is not exposed to the user, it is an implementation detail, and what I want to do now is to hide it even in the classes that I use internally, so the complexity is contained even more.

What I basically wanted to know is whether it is safe to static cast void*[] to MyClass*[] (assuming MyClass pointers is what is stored there). I've got an answer that it is not safe(and requires reinterpret_cast), I'm happy with that answer and not going to look further. But now I have to protect my design decisions of the whole library for some reason.

Damn, telegram doesn't allow me to write ** twice even in code blocks

Damn, telegram doesn't allow me to write ** twice even in code blocks

It's a shame that telegram web has syntax highlighting and it's not available on the official client yet.

It's so foolish "to rely on the standard" because C++ is defined by both the standard and the real world, as I already said

Well following the standard has the pro of being multiplatform, it also has the con of making you completely paranoid

Well following the standard has the pro of being multiplatform, it also has the con of making you completely paranoid

I agree with following standard I just don't agree that it's the only truth and C++ is only limited by it.

I agree with following standard I just don't agree that it's the only truth and C++ is only limited by it.

Microsoft learned that dire truth when they had to reimplement templates from scratch in 2016

Microsoft learned that dire truth when they had to reimplement templates from scratch in 2016

(also they were not quite following the standard already with that so it was a good sting in their asses)

We live in a world where compiler developers can understand the standard the way they want to The standard is the good thing, yet it's not the only way how C++ is defined There is no C++ implementation that 100% follows the standard because it's technically impossible

So when writing code you need to use your knowledge about the world and not blindly believe the standard

We live in a world where compiler developers can understand the standard the way they want to The standard is the good thing, yet it's not the only way how C++ is defined There is no C++ implementation that 100% follows the standard because it's technically impossible

Yes, I was just poking some fun at Microsoft doing something differently from literally anyone else that changed the semantics and usage of wide parts of the language. (when I mean everyone else, I mean gcc, clang, IBM, cray, digital mars, Nvidia at the very least)

Yes, I was just poking some fun at Microsoft doing something differently from literally anyone else that changed the semantics and usage of wide parts of the language. (when I mean everyone else, I mean gcc, clang, IBM, cray, digital mars, Nvidia at the very least)

Yeah, they have "interesting" stuff going on

Yet in the last years they are pioneers of the C++ standard

Yet in the last years they are pioneers of the C++ standard

Well, they had to rewrite a lot of the compiler so they made it more maintanable

Yet in the last years they are pioneers of the C++ standard

ever worked with c++/clr? it's one of the biggest messes I've seen

ever worked with c++/clr? it's one of the biggest messes I've seen

I used it, I hate it, but it had a very significant performance benefit compared to regular MS VC++

I used it, I hate it, but it had a very significant performance benefit compared to regular MS VC++

what on earth... how..

ever worked with c++/clr? it's one of the biggest messes I've seen

C++/CLI is not C++

what on earth... how..

The runtime would run the optimizer upon launching the code again, with native specific optimizations on, giving around 12% of improvement on several of my test loads, and no loss on other loads

The runtime would run the optimizer upon launching the code again, with native specific optimizations on, giving around 12% of improvement on several of my test loads, and no loss on other loads

i wonder if you could've achieved the same thing using any other language with a tracing jit, luajit for example

The runtime would run the optimizer upon launching the code again, with native specific optimizations on, giving around 12% of improvement on several of my test loads, and no loss on other loads

Well first of all MSVC optimizer is a joke Then, you should've compiled with something like -march=native

Well first of all MSVC optimizer is a joke Then, you should've compiled with something like -march=native

I said "compared to normal MSVC"

i wonder if you could've achieved the same thing using any other language with a tracing jit, luajit for example

did you benefit from the features CLI provides at all?

Well first of all MSVC optimizer is a joke Then, you should've compiled with something like -march=native

Then even if it'd have a performance benefits, they wouldn't be so drastic

i wonder if you could've achieved the same thing using any other language with a tracing jit, luajit for example

Sure, sometimes jit programs are faster than their native counterparts But it's more like luck than the pattern

did you benefit from the features CLI provides at all?

Not much, just in the easier interfacing with C#

Well first of all MSVC optimizer is a joke Then, you should've compiled with something like -march=native

I didn't have control on what servers the code would run on, and it would run on servers that I didn't know the architecture of

Well first of all MSVC optimizer is a joke Then, you should've compiled with something like -march=native

fun thing, their optimizations (even when talking about architecture independent things) are much better when targeting x64 than when targeting x86. i wonder what kind of crap is going on under the hood for such a thing to occur

I didn't have control on what servers the code would run on, and it would run on servers that I didn't know the architecture of

That's sad

fun thing, their optimizations (even when talking about architecture independent things) are much better when targeting x64 than when targeting x86. i wonder what kind of crap is going on under the hood for such a thing to occur

https://gcc.godbolt.org/z/sWdPxo861

fun thing, their optimizations (even when talking about architecture independent things) are much better when targeting x64 than when targeting x86. i wonder what kind of crap is going on under the hood for such a thing to occur

Well, if I were to build a cross platform software, I'd just choose clang for all possible platforms and lld

https://gcc.godbolt.org/z/sWdPxo861

there's no reason for something like this to ever happen

Hello, I want to using native api on Windows. I've tried phnt headers but it shows error. Can you guys suggest me a native api for windows header?

How can I access global variable from function only in the same module without changing the signature of the function.

How can I access global variable from function only in the same module without changing the signature of the function.

May you give an example. I'm not sure I understand what you mean.

May you give an example. I'm not sure I understand what you mean.

int foo; foo = 5; Int bar() { return foo;} bar() // return 5 foo = 6; bar() // return 6 // This kind of functionality

You cannot chenge wariable out of function scope.... C++ is not a script language. It runs in function body only.

May you give an example. I'm not sure I understand what you mean.

Solved with extern🌹 C not cpp

If you will wrap first bar and second bar with the function you will get an axpected result.

Extern you need when you have more then one cpp file or more then one module. So you will told to the compiler that you will bring the address of the variable during the linking and not at the compile time.

How can I access global variable from function only in the same module without changing the signature of the function.

If you are talking about C++20 modules, then there is some module linkage. https://vector-of-bool.github.io/2019/10/07/modules-3.html If by "module" you refer to a translation unit, you can mark the global variable as static or put it in an unnamed namespace (it should be declared in *.cpp file, not in *.h), then it will be available only in the translation unit where you declared it (and if you declare it in different translation units, each unit would have its own instance of the variable). Note that such variables are tricky to manage, and any global variables are in general.

You cannot chenge wariable out of function scope.... C++ is not a script language. It runs in function body only.

Ohk . it is not in c In c only const variable cannot be change when it's in romb