Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
Search - "std::function"
-
- Let's write some code to check for memory leaks
- Oh shit, memory is leaking like crazy
- In fact the program crashes within 10 minutes
*Some hours of debugging and not finding the cause later*
- Starts thinking about the worse
- Hell yeah, the memory leak is caused by the code that checks for memory leaks. But fucking how
- Finds out the leak is caused by the implementation of the std C lib
- In the fucking printf() function
- Proceeds to cry5 -
I am currently in a bit of a (well-deserved) lull at work, both of my projects are finishing up/ finished, so tomorrow should be pretty light, as the latter half of today was.
And I have really gotten interested in the HTTP protocol. It's so interesting learning how it all works under the hood.
So I think I'm going to be researching/ messing around with creating a cpp project that essentially implements cURL from the ground up, creating sockets, reading from them, parsing the HTTP requests... all that. I don't expect to actually get it done, but it should be an immense learning experience. I have a clear goal: implement this function:
std::string get(const std::string&);
Once I'm able to just GET as simple as that, I know I have achieved my goal!3 -
I started my internship at the end of the year..
Fuck my ass!!! This code I have to work with is a huge pile of shit.
The code base I need to work with is around 40k LOC. It is a mixture of C++, C, Java, Python, Bash and I think I saw some lonely js files around.
A list of awesome parts:
- Paths are hard coded.
- Redundant code everywhere
- No documentation or inline comments available
Most of the comments in the code are just old code that is not used anymore. But the cherry on the turd is the class that should provide all kind of useful functions in my daily routine. About ninety percent of the functions have the same description or nothing. Sometimes a function name says "readSomethingFromSomewhere" but instead it writes something to a file. It is really confusing and I need to check everything twice instead of rely on what the function name promises.
I have also learned why copy paste isn't that good. The brief descriptions of every method in a files are always the same.
getName() - Description: Fork child process
getIp() - Description: Fork child process
getIpv6() - Description: Fork child process.
Surprise: None of these functions forks a child process. :D
Another awesome feature is the thing that they store up to five different versions of libraries. Everyone with slight modifications but no hint which one you need to use. Sometimes it is the newest, sometimes the oldest which is running in production. Another case of try and error.
Oh and my dev machine is a potato with a power supply and a fan. I started with NetBeans and every time I compiled the code it sounds like the machine wants to lift off and leave for a better place. (At this point I switched to Emacs and everything runs smoothly now)
At first I thought that I'm just not that good at coding and understanding a big project from scratch but some colleagues have the same problem. The whole system is very inflexible and it is all about "std::cout"-debugging to check if your changes do what you want them to do.
Currently I'm just trying to fix this mess to make the life for the next student or employee easier. The first month was just frustrating as hell. I need to ask so many questions and most of the time the answer was "I don't know, haven't touched this code in years". Needless to say that my progress isn't that awesome but at least I get a nice payment for 20 hours of work a week.2 -
Any other language: Hey fuckface, you can't name this variable by a single letter, tf is wrong with you? use some descriptive shit.
Golang: lmao fuck u
I really find it interesting how we use short variable names for items in golang. Kinda makes sense when you think of it. Most of these items come up in short methods for which the mental model lets you know and remember what you are doing, they even make sense when going through the std lib in which that shit is all over the place. YET years of going by other languages has made me squint my eyes a bit in frustration every time I see it.
Say for example that a function is implementing io.Writer. What would you call the method parameter? you could argue that writer would be sensible since it has it in the signature, but what about when the io.Writer in itself is a file or a socket or whatever? writer would be funny or strange? nah fuck it just w, it makes sense, but x wouldn't. I find these points to make sense even if i don't like them.
Would, now, this practice be acceptable in C? you are supposed to write the same modular code with C in which you compose large functionality in separated units of code, yet I am sure this practice of single name variables is something that C engineers dislike greatly.
Are go devs just doing this out of blind love for their preference in languages? and how would this work if mfkers add generics to go(I hope not, Go is simple enough to understand in order to extend functionality through the empty interface, but that is a preference of mine as well)
The more I use Go the more I like it to be honest, I think the code looks ugly syntactically, but that is subjective as all hell and based on my constant preference for a language to look like Ruby, which even though it might not be everyone's cup of tea it remains to my eyes as the most beautiful language in existence, again, an obvious personal preference.18 -
The crazy shenanigans you can do with C++ standard libs are fascinating.
Like implementig multithreading with just a foreach, and bindings which can make member function pointers to simple function pointers, and placeholders in bindings. Also lambda functions are cool.
Something between the lines:
my_crazy_class *tmp = new my_crazy_class(...);
std::vector<type> my_array = .....;
std::for_each(std::execution::par,my_array.begin(),my_array.end(),
[&](type in){
auto fn = std::bind( &my_crazy_class::my_crazy_fnc,*tmp,_1,random_static_value);
return fn(in);
});
ps:
It's pretty much pseudocode, and please don't do things like this, it's bad for your mental health.
pps:
I need to learn how to use this tools wisely. -
Why the pointer of member function so special.
trying make a non-template class that like std::function but simpler:
1. wrap static function pointer.
2. wrap class member function pointer and the class object pointer.
3. WITHOUT ALLOCATE MEMORY.
I can't store member function pointer because the size is undefined?! -
Just realized a member function pointer can be a template parameter as non-type, gonna try to use it do no dynamic memory allocation trick with std::function.
-
How deep do you go when trying to find a solution?
I have a need for combinations of items. I have used built in functions in Python for this. When I first used those I wanted to learn how they worked internally. I read through the source and thought that was cool. I don't think I really understood that code very well.
Now I need the same solution in C++. There is not a prebuilt combinations function in C++. There is a prebuilt verion of next_permutation. I can build upon that to make my combinations code. However, I am in the middle of trying to make something work. So I found this nice SO question:
https://stackoverflow.com/questions...
The code I ended up using:
template<class RandIt, class Compare>
bool next_k_permutation(RandIt first, RandIt mid, RandIt last, Compare comp){
std::sort(mid, last, std::bind(comp, std::placeholders::_2, std::placeholders::_1));
return std::next_permutation(first, last, comp);
}
template<class BiDiIt, class Compare>
bool next_combination(BiDiIt first, BiDiIt mid, BiDiIt last, Compare comp)
{
bool result;
do{
result = next_k_permutation(first, mid, last, comp);
}while(std::adjacent_find(first, mid, std::bind(comp, std::placeholders::_2, std::placeholders::_1)) != mid);
return result;
}
I am mostly able to figure out what is going on with the templates. I still am not understanding the basic algo behind permutations.
Our data set is tiny. 4 items max. So efficiency isn't really a big issue here.
How long do you spend learning how it works versus just finding a solution for the task at hand?
In general I need to spend more time learning different kinds of algorithms. So I should probably add permutations to that list of ones to study.1