Do not use rand and time

It's recommended by reference

So notstd::single_view can be creates a new object, but what if I want to create it from already existing object? Only copy?

that is a good point, working on that

It's recommended by reference

no. use C++ random number engines

why?

Rand has bad distribution Time is just C API, meh

There are <random> and <chrono>

It's recommended by reference

What reference?

but if i just want to get "random" number, not for crypto... is it bad to use rand? do i need some other lib in c?

but if i just want to get "random" number, not for crypto... is it bad to use rand? do i need some other lib in c?

In C++ there are <random>

indeed

that is a good point, working on that

You can use smart pointer "idiom" Constructors for creating single_view from existing object And make_single_view template function to create object with args

What reference?

https://en.cppreference.com/mwiki/index.php?title=cpp/numeric/random/rand

You can use smart pointer "idiom" Constructors for creating single_view from existing object And make_single_view template function to create object with args

ye this is excellent. thanks. i was wondering how i would handle the different cases

But I dunno how to combine storing a reference/pointer to existing object or newly created object

ye this is excellent. thanks. i was wondering how i would handle the different cases

Kinda std::variant<T*, T>

But I don't think that you even should provide this feature with make_single_view

https://en.cppreference.com/mwiki/index.php?title=cpp/numeric/random/rand

It doesn't recommend anything It's just a reference

https://en.cppreference.com/mwiki/index.php?title=cpp/numeric/random/rand

Every sane C++ developer recommends you not to use rand

But I don't think that you even should provide this feature with make_single_view

I mean view is semantically meant to be created from an existing object Like std::string_view, std::span

But I dunno how to combine storing a reference/pointer to existing object or newly created object

i can keep two pointers, one unique_ptr and one normal pointer. for objects created with make_single_view unique_ptr will own the object and the normal pointer will contain result of unique_ptr::get(). for pointers directly passed to single_view the smart pointer will remain nullptr.

i can keep two pointers, one unique_ptr and one normal pointer. for objects created with make_single_view unique_ptr will own the object and the normal pointer will contain result of unique_ptr::get(). for pointers directly passed to single_view the smart pointer will remain nullptr.

Nah that's not good

I mean view is semantically meant to be created from an existing object Like std::string_view, std::span

the std::single_view actually requires T to be copy constructible. there is also std::iota_view

std::string res = "0000000000"; Error: no viable conversion from 'const char [11]' to 'std::string' (aka 'basic_string<char>') WTF?

std::string res = "0000000000"; Error: no viable conversion from 'const char [11]' to 'std::string' (aka 'basic_string<char>') WTF?

probably mistake before this line

probably mistake before this line

The line before: std::string randidgen() {

std::string res = "0000000000"; Error: no viable conversion from 'const char [11]' to 'std::string' (aka 'basic_string<char>') WTF?

Paste the whole code

Paste the whole code

Solved already, it was a problem with my IDE

The line before: std::string randidgen() {

:/

Solved already, it was a problem with my IDE

oh

i can keep two pointers, one unique_ptr and one normal pointer. for objects created with make_single_view unique_ptr will own the object and the normal pointer will contain result of unique_ptr::get(). for pointers directly passed to single_view the smart pointer will remain nullptr.

I don't like it, because there is unnecessary memory allocation

I don't like it, because there is unnecessary memory allocation

yes. but i don't see any other way for me to simultaneously work on objects i own and not own

one more way might be

Kinda std::variant<T*, T>

What about this?

reading

What about this?

1) it's C++17, i wanted C++11 compatibility. 2) i think that will get overcomplicated with std::holds_alternative once every 2 lines in the iterator template

C++ 20 is out or soon? My head phones need replacing as the cable is loose inside

C++ 20 is out or soon? My head phones need replacing as the cable is loose inside

What

1) it's C++17, i wanted C++11 compatibility. 2) i think that will get overcomplicated with std::holds_alternative once every 2 lines in the iterator template

actually 2 might not be the case. only the begin and end functions of the view would need replacement. but still it's C++17 only

https://www.modernescpp.com/index.php/c-20-an-overview

What

I think he wants write a cable for his headphones with C++20

I was listening to a podcast. But my head phones are broken. So I thought I could ask here if it's releaed

I think he wants write a cable for his headphones with C++20

Oh interesting carry on

C++ 20 is out or soon? My head phones need replacing as the cable is loose inside

lol what

And does clang support it

C++ 20 is out or soon? My head phones need replacing as the cable is loose inside

C++20 standard is closed afaik but none of the compilers supports it fully

Ty

And does clang support it

https://en.cppreference.com/w/cpp/compiler_support

tyvm

https://en.cppreference.com/w/cpp/compiler_support

GCC has really good support for C++20

I wonder why other compilers are so slow...

Because they implement more

Because they implement more

What do you mean?

What do you mean?

GCC doesn't have comparable features to others

GCC doesn't have comparable features to others

What do you mean?😂

The only thing clang hasn't implemented is stack probing on arm64 from what I'm aware

Otherwise GCC doesn't have half of clang's sanitizers, no thinlto and is missing a lot more when it comes to not breaking code

Just out of interest, is it possible to generate random data out of undefined behavior?

Just out of interest, is it possible to generate random data out of undefined behavior?

no. then undefined behaviour will be defined as "produces random data". this is a contradiction since undefined behaviour is by definition undefined

no. then undefined behaviour will be defined as "produces random data". this is a contradiction since undefined behaviour is by definition undefined

Okay.

So notstd::single_view can be creates a new object, but what if I want to create it from already existing object? Only copy?

technically an user can use range/view functions from the STL to create a range that doesn't own the object. then they can pass that range to my single_view. they can also pass a pointer to my single_view

technically an user can use range/view functions from the STL to create a range that doesn't own the object. then they can pass that range to my single_view. they can also pass a pointer to my single_view

But that would not reliabily be random data, right? And it could cause a segmentation fault.

But that would not reliabily be random data, right? And it could cause a segmentation fault.

i am talking about my single_view class :/

oh

nvm then 😂

But that would not reliabily be random data, right? And it could cause a segmentation fault.

i mean it can cause segmentation fault. but so can std::string_view and other STL views that don't own the object

Hi, can someone help me? I need to create copying comstructor to this class

template<typename T> class SSL { struct Node { T data; Node* next; }; Node* head = nullptr; public: // ... };

That creating deep copy of the transmitted object

Can someone help with this?

Hi, can someone help me? I need to create copying comstructor to this class

Check out cppreference site to see about copy constructors

I check thisand really dont know how to make deep copying constructor for this class...

Help me to make flowchart and pseudocode for my mini project

i dont understand gmp export. is mpz_export(buffer, NULL, 1,1, 0, 0, mydatampz); correct?

size of mydatampz is 1024 but buffer use 128 slot of 1024

i dont understand gmp export. is mpz_export(buffer, NULL, 1,1, 0, 0, mydatampz); correct?

https://gmplib.org/manual/Integer-Import-and-Export#Integer-Import-and-Export

Hi, any idea why "thisсlient->user" results in "stray '\201' in program"?

solved, с != c

what happens if i only use 128 slots of an array of 1024 ?. does the rest of the memory remain accessible?

Yes

what happens if i only use 128 slots of an array of 1024 ?. does the rest of the memory remain accessible?

if you allocate 1024 slots, you have all the slots assigned to that array and only that array will have such slots reserved if your code does not actually use all of them, it is a choice you make at a logic level and other components of the program will not be able to benefit from that memory, that is actually wasted it is your responsibility to decide wether such a waste is convenient and acceptable or not

what happens if i only use 128 slots of an array of 1024 ?. does the rest of the memory remain accessible?

otherwise you have to reallocate the array in a smaller memory region, but reallocations are not an efficient process