It seems the only difference betweens binary or text is how to translate \n to end of line. So it seems not a good ideal to detect whether a stream is binary or text

+1 to klimi shared_ptr should be used to represent shared ownership (when multiple places control lifetime of an object). If you don't have shared ownership it's better to default to unique_ptr. Also shared_ptr can be constructed from moved unique_ptr, so if you need shared ownership later (e.g. you making a factory and don't know how the created objects will be used), then unique_ptr is a good default also

I replaced all my raw pointers with shared_ptr's and I'm using only make_shared, but something strange is happening... After all variables going out of scoped, I'm having 84 constructor calls (expected), but only 14 destructor calls https://pastebin.com/Zh7Yv9YL

I think unique_ptr wouldn't work because they are all tree nodes and I need to access them from many different function, with ownership, like insert, remove, rotate

Hi,when installing a mingw on w7 thes are facing a issue about download faild any solution?

This is wrong. You are using a shared_ptr to represent both children and parent node relationship. Now if a child node points to the parent and the parent points to the child using shared pointers,think how these objects can be freed. They can't because the reference count for neither the parent nor the child nodes will get reduced. This is a situation where you should use a raw pointer. The relationship from parent to child should use a shared_ptr/unique_ptr. The relationship from child back to parent should be a raw pointer.

This is wrong. You are using a shared_ptr to represent both children and parent node relationship. Now if a child node points to the parent and the parent points to the child using shared pointers,think how these objects can be freed. They can't because the reference count for neither the parent nor the child nodes will get reduced. This is a situation where you should use a raw pointer. The relationship from parent to child should use a shared_ptr/unique_ptr. The relationship from child back to parent should be a raw pointer.

"The relationship from child back to parent should be a raw pointer" Could I use a weak_ptr instead?

I can debug CMake with gdb?

Okay, thanks 👍 So, I'm making a small library with which (simplifying) I can call a function and execute a program/command. Example: auto r = Execute("echo hello"); r.stdOut will be hello. I have the basic setup with CMake. I managed to add a testing folder and make it work but I have a little problem. For testing I would like not to rely on the OS programs, but rely on some programs written by me. I want that every time I generate the testing executable other executables are generated, one from each .cpp file in a determined folder. I tried with a foreach but it kept giving me errors, how should I do this?

no. only one benchmark

Good afternoon again. If possible I would like to ask. I don't understand why I have a problem with the file format. If all I do is just insert pieces of working code? Is this a problem in the compiler itself or am I doing something wrong. I will be grateful for any help https://onlinegdb.com/LCoI4ShcX

Replace do...while with a while loop, "read" returns bytes not total amount of items, so line 51: if (read == 0 && !feof(file)) { printf("File format incorrect.\n"); return 1; }

So I finally got my hands on it. I see comparable results on both gcc and clang (differences fluctuate from run to run, and are very small), so they run basically the same in this case https://quick-bench.com/q/duWjGxfX8F1Zog0Yi2RnGV0CaAc What is different with your benchmark?

is there some built-in function in c++ to determine if a filesystem::path is image?

is there some built-in function in c++ to determine if a filesystem::path is image?

There is not. That would require opening the file for examination. You could the analyse the start of the file to see if it could be a known image format, or you could analyse the whole file to see if it is a valid image of that format

There are dozens upon dozens of image formats and standardizing this introspection would definitely require insane work

Good evening again! I faced that I do not know how it is possible to compare in an element of structure. For example, car numbers АВ123НР, ВА445СН and the like. Identify only those numbers in which consist of different digits. For example, after work I need the program output AB123HP. I will be grateful for any help or a link to the necessary information because I can not find anything like that

First of all, you will need to do some analysis of the character string that makes up the car number. C++ comes with the guarantee that any UTF-8 character with a codepoint below 128 will be correctly stored in a string. It also guarantees that the C function isdigit will work, allowing you to extract each digit in something of type char. Things of type char can be compared for equality, and it will require exactly 3 comparisons to do your thing

This is of course only valid if your vendor distributes C and C++ as a unit. If it doesn't, it is perfectly valid for the character set to have various representations for every digit, making comparisons unsuited for your task

This is of course only valid if your vendor distributes C and C++ as a unit. If it doesn't, it is perfectly valid for the character set to have various representations for every digit, making comparisons unsuited for your task

Sorry to bother you again, but if, for example, my car number consists exclusively of numbers. Will the task become less difficult?

It kinda depends, are you in C or in C++?

you may want to: - count how many digits you have - allocate memory for every digit - store every digit - compare then you can also compare in place, but it may seem less intuitive upon reading it

What field do you guys find it exciting in C++?

I have seen all sorts of fields using C++, it is a very flexible language and can do almost anything. I am personally passionate about database engineering and distributed systems development. It is widely applicable and a very complex field with lots of resources to delve in and learn from

Can you tell me what are those fields? I know only two main ones : Gaming, system programming

the compiler do some optimization，when the loop is fixing number，unroll

I didn't know there was such field called "Database engineering"

Meaning writing database engines, query protocols and ensuring constraints on both of those

For example, ensuring data consistency, data access, enforcement of linearisation or serialization constraints etc etc

What is query protocol?

Serialization constraints? Like what constraints are applied when converting some disk data to SQL query result?

The most common constraint is ordering of operation, which gives a strict ordering on which several transactions appear to have been performed. This is of crucial importance to ensure transactions do not "overlap" and stop being isolated

A question that always occurs to me: what happens when two or more transactions are received at the same time, like which one will be chosen?

Use gets function

Use gets function

> iostream.h

How to solve this problem? Node depends on Connection, and Connection depends on Node. When it is a function you can just declare the contract and later you write the implementation class Node { public: int data; vector<Connection> neighbors; Node(int data) : data(data) {} }; class Connection { public: Node *node; int distance; Connection(Node *node, int distance) : node(node), distance(distance) {} };

How to solve this problem? Node depends on Connection, and Connection depends on Node. When it is a function you can just declare the contract and later you write the implementation class Node { public: int data; vector<Connection> neighbors; Node(int data) : data(data) {} }; class Connection { public: Node *node; int distance; Connection(Node *node, int distance) : node(node), distance(distance) {} };

#include <bits/stdc++.h> using namespace std; int main() { int s=0,n; cin >> n; for(int i=1;i<=n;i++) { s=s+pow(-1,i)*i; } cout<<s<<endl; return 0; }

If i%2 == 1 then pow(-1,i) is -1 else it is 1.

#include <bits/stdc++.h> using namespace std; int main() { long s=0,n; scanf("%ld",&n); for(int i=1;i<=n;i++) { if(i%2==0) s=s+i; else s=s-i; } cout<<s<<endl; return 0; }

you can define class prototype just as function at the start of the program. Like class connection; class node; and then you can define them anywhere so compiler will know where to find them.

#include <bits/stdc++.h> using namespace std; int main() { long s=0,n; scanf("%ld",&n); for(int i=1;i<=n;i++) { if(i%2==0) s=s+i; else s=s-i; } cout<<s<<endl; return 0; }

STOP USING bits/stdc++.h header, That's not portable

where can I practice my c skills

hackerrank