C++ - Reddit

Layout of C++ function member pointers for MS/Itanium ABI https://lief.quarkslab.com/blog/2021-04-08-profiling-cpp-code-with-frida-part2/ https://redd.it/mqju21 @r_cpp

cannot return string value Code: \#include <iostream> \#include <string> &#x200B; int redactedForPost() { std::cout << "REDACTED"; std::string input1; std::cin >> input1; &#x200B; return input1; } &#x200B; // Main part of script int main() { return 0; } The code is unfinished but at the: return input1; line it says I cannot return a local variable. https://redd.it/mqefq3 @r_cpp

Invalid arguments ' Hi, this is a small problem i hope. I'm getting this when trying to add a pointer to my function call, basically just after adding the & operator to my string params. The vectors seem to be fine with this &, but i removed them anyway. int pairMatchDecider(string (&arrDeckOne), string (&arrDeckTwo), vector<string>deckOne, vector<string>deckTwo) { &#x200B; pairMatchDecider(arrDeckOne, arrDeckTwo, deckOne, deckTwo); https://redd.it/mqf519 @r_cpp

Complex C++ Compilation from the Command Line https://unrealistic.dev/posts/introduction-to-c-compilation-on-the-command-line https://redd.it/mq58qv @r_cpp

Qt Creator 4.15: New CMake Features https://www.qt.io/blog/qt-creator-4.15-cmake-new-features https://redd.it/mpy29p @r_cpp

Polymorphic Usage Of Non - Polymorphic Class https://m-peko.github.io/craft-cpp/posts/polymorphic-usage-of-non-polymorphic-class/ https://redd.it/mpvoya @r_cpp

Redesigning Legacy Systems - Peter Muldoon - ACCU 2021 A brand new talk for ACCU 2021 - some place we have all been, how to successfully rewrite and replace a Legacy System. Feedback welcomed. Redesigning Legacy Systems https://redd.it/mpmf7d @r_cpp

std::forward_list vs std::vector compiler optimizations While benchmarking some specific class in my code I realized that using `std::forward_list` instead of `std::vector` for its backing storage resulted in greater performance when storing a small number of elements. This sounded counter-intuitive and I watched enough Chandler Carruth's talks to know that `std::list`/`std::forward_list` is almost always the worst choice of container when speed is concerned. I tried a few things to improve the performance using `std::vector`, including reserving some memory beforehand but it was still beaten by `std::forward_list`. So I opened up compiler explorer and tried to figure out what was actually going on. I quickly realized that when using `std::forward_list` the code was much better optimized with around half the instruction count compared to the `std::vector` version and almost no function calls. I guess this explain the performance difference I observed. I then tried to get the smallest possible piece of code that would highlight this difference and I ended up with [this](https://godbolt.org/z/WxrdG63rG). It's just a `push_back(12)`/`push_front(12)` and then returning `back()`/`front()`. You can see that the `std::forward_list` version is completely optimized and simply returns 12 while the `std::vector` version performs memory allocation/deallocation while still unconditionally returning 12. Is there a fundamental reason why `std::forward_list` gets better optimized than `std::vector`? Or is it just due to inconsistent optimizations? Also, since memory allocations are kept with `std::vector` while the return value is harcoded I guess that the compiler assumes that calling `new`/`delete` may have side effects. But if it's the case, why assuming that there aren't side effects with `std::forward_list` and removing the calls? https://redd.it/mpiurp @r_cpp

I'm working on a template class for a matrix container. I've just implemented a method to calculate the determinant through template recursion. Any thoughts? https://github.com/cristi1990an/Matrix https://redd.it/mpkb99 @r_cpp

C++ course for intermediate to advanced programmer I've been away from C++ for a few years, but I've coded for a long time. I've taken a few courses on Udemy... skimmed most of them as the content is for beginners. Is there a good intermediate to advanced course on C++ that you recommend? Ideally something w/projects and that uses *nix? https://redd.it/mpi34r @r_cpp

A new face for the Qt Project https://www.qt.io/blog/a-new-face-for-the-qt-project https://redd.it/mphcox @r_cpp

cmake-init - The missing CMake project initializer https://github.com/friendlyanon/cmake-init https://redd.it/mowaqg @r_cpp

Strong Types in C++: A Concrete Example https://www.cppstories.com/2021/strong-types-pesel/ https://redd.it/mp8zuy @r_cpp

Environment for new C++ project Hi all, I'm about to embark on a new C++ project fairly soon and I'm looking for some advice on build systems and tooling. My professional experience to date has always been working on existing projects with custom build tools, so I'm not very familiar with the defacto standards for real projects. I'm looking to compile application software for windows and potentially down the line Linux as well. One of the requirements is that all of the tooling be self hosted. I am looking at using clang as my primary compiler, with ninja build scripts generated by either meson or cmake. 3rd party libraries like Qt, boost, etc including binary only redistributables I'm intending to manage with Conan, using an Artifacfory CE server that we host internally. Unit testing will likely be done through Boost.Test, although I'm not too worried about the test framework at this stage. One thing I would really like to be able to support is having a package in Conan/Artifacfory that is a bundled up version of the compiler I am intending to target, and having the build scripts be able to fetch that compiler and store it in a Conan cache on disk, then call to it when I want to compile. Ideally this would mean that everyone is using the same version of the same compiler, and wouldn't have to install the compiler themselves. Does anyone know if this is possible with my proposed environment? If so, are there any examples that I could be linked to or that you could share with me? https://redd.it/mp5xln @r_cpp

NFT backed implementation of std::unique_ptr https://github.com/zhuowei/nft_ptr https://redd.it/mp5kss @r_cpp

Pass by value doubts Hi i am a new member here and was wondering if someone could help me with a problem in c++ the question is from hackerearth https://www.hackerearth.com/practice/codemonk/ i am trying to send each element of the array to a function where in the function will compare that element with all the elements succeeding it in the array and if the value of the passed parameter is more in any comparison then i am incrementing the count and finally returning the count for each array element which adds up. but i am facing this problems can anyone tell me what is wrong with the syntax for now ? &#x200B; my code \#include <iostream> using namespace std ; int fun (int num,int i,int j,int n,int l[n\]) { int inv=0; for (int r=i;r<n;r++) for(int c=j+1;j<n;j++) { if(num>l[r\][c\]) inv++; } return inv; } int main(){ int t; cin>>t; while(t--) { int n; cin>>n; int f=0; int a[n\][n\]; for(int i=0;i<n;i++) {for(int j=0;j<n;j++) { cin>>a[i\][j\]; }} {for(int i=0;i<n;i++) {for(int j=0;j<n;j++) int x=a[i\][j\]; f=f+ fun(x,i,j,n,a); }} cout<<f; } } error message &#x200B; 5:43: error: use of parameter outside function body before ‘\]’ token In function ‘int fun(...)’: 8:12: error: ‘i’ was not declared in this scope 8:16: error: ‘n’ was not declared in this scope 9:11: error: ‘j’ was not declared in this scope 10:9: error: ‘num’ was not declared in this scope 10:13: error: ‘l’ was not declared in this scope In function ‘int main()’: 41:11: warning: unused variable ‘x’ [-Wunused-variable\] 42:18: error: ‘x’ was not declared in this scope 42:22: error: ‘j’ was not declared in this scope https://redd.it/mp3lku @r_cpp

What do you think is faster for batch-processing a lot of "double-type" arithmetic? Without saying, this is for a per-thread basis. The whole batch is of course split into threaded workloads. Option a) Use the avx featureset to process four doubles at once ... about 12 avx calculations. Maybe a few less if using FMA Option b) Dynamically cast all 16 doubles to floats Then use the avx featureset to process eight floats at once ... about 6 avx calculations. Maybe a few less if using FMA Dynamically cast all 16 floats back to doubles This specific operation does not need double-precision. Other less frequent operations run on these variables do. So just switching to floats for good is not an option. This really is a question if roughly halving the calculation time is worth dynamically casting all variables TWICE... Which option do you think runs faster? https://redd.it/motehs @r_cpp

Code style tool Is there a good tool for Eclipse that forces a code style/lint? https://redd.it/mor5fb @r_cpp

Kris Jusiak - Online AVR/Arduino Developement with C++20 - Meeting Embedded 2020 https://www.youtube.com/watch?v=MskKTIdW45c https://redd.it/moqs44 @r_cpp

C++11/14 at Scale - What Have We Learned? - ACCU 2021 https://www.youtube.com/watch?v=H8wzuvynV78 https://redd.it/mood6s @r_cpp