@programminginc - страница 2334

Ammar

Hey Ajay, take a look to std::fill_n. It is probably what are you looking for.

Ajay

Hey Ajay, take a look to std::fill_n. It is probably what are you looking for.

thanx buddy, I know that

Ajay

I often get segmentation fault in my cpp code, debugging is a way but are there any other ways to tackle with this problem or ways that can help to find at which line did the exception occurred?

Ammar

I often get segmentation fault in my cpp code, debugging is a way but are there any other ways to tackle with this problem or ways that can help to find at which line did the exception occurred?

Compile your code with flag -ggdb3, and run it with gdb. It will tell you the line of code which it gets SIGSEGV.

Ammar

Valgrind should also be helpful, it also tracks invalid read/write and memory leaks.

noop

I often get segmentation fault in my cpp code, debugging is a way but are there any other ways to tackle with this problem or ways that can help to find at which line did the exception occurred?

Use smart pointers Use code static analysis tools

Dima

Use smart pointers Use code static analysis tools

smart pointers are making it way much easier:/

Ammar

Best tools for debugging: gdb, valgrind, radare2

Ammar

Also use gef for your gdb. It makes you easier to see instruction, registers and stack.

Mar!o

I often get segmentation fault in my cpp code, debugging is a way but are there any other ways to tackle with this problem or ways that can help to find at which line did the exception occurred?

and double check your double pointers

Mar!o

my classic was: void free_stuff(stuff **restrict const stuff) { free(stuff); <— should be *stuff *stuff = NULL; }

Anonymous

I'm running a function on a thread, and it contains a gets that asks for input. Now, after starting this thread, I decide that I don't need the input anymore. Is there any way I can "close" the gets prompt? Like supplying it with some random data?

Ammar

I'm running a function on a thread, and it contains a gets that asks for input. Now, after starting this thread, I decide that I don't need the input anymore. Is there any way I can "close" the gets prompt? Like supplying it with some random data?

It is probably possible to supply stdin via /proc/$pid/fd/0

Ammar

I am unsure btw

Anonymous

That's the problem. It cannot wait. I won't know if the input comes first, or the decision that I don't need the input does.

Ammar

Just thought that it was Linux.

Ammar

I wonder what is the purpose of that folder.

Ammar

I see, thanks.

Ammar

Try this one btw https://stackoverflow.com/questions/38329044/closing-then-opening-standard-input-in-c-under-linux

Anonymous

Well, I'm trying to run a chat server, that constantly asks for user input, and prints a message immediately when it's received. The receiver function also runs on a thread. I'll try to think of some other way then.

Ammar

I haven't tried but this eww https://serverfault.com/questions/188936/writing-to-stdin-of-background-process

Ajay

Compile your code with flag -ggdb3, and run it with gdb. It will tell you the line of code which it gets SIGSEGV.

I just started using gdb so I wrote a simple hello world program, and ran gdb, after gdb starts, run command should work right? and should output on the console right?

Ammar

I just started using gdb so I wrote a simple hello world program, and ran gdb, after gdb starts, run command should work right? and should output on the console right?

Yes, sure.

Ajay

Yes, sure.

it kinda stops there and does nothing

Ajay

like this, IDK why it isn't outputting anything.

Ammar

Ok, I agree with you.

Ammar

like this, IDK why it isn't outputting anything.

Looks it is waiting for something.

Ajay

Looks it is waiting for something.

I didn't provide any breakpoint

Ammar

Try to send SIGINT and check the current instruction.

Ammar

CTRL + C x/10i $rip

Ammar

I assume the arch is x86_64

Ludovic 'Archivist'

It is the same amount of skipped items for any hash, it is an increasingly large amount until you have failed more than 3000 probings and then it flips back to linear probing

Ajay

it is simple hello world, I was just trying it for my previous code but it kinda started waiting, so I wrote a simple hello world to check if it still waits and see it still is waiting. IDK why?

Anonymous

If you want the details, the receiving function stops as soon as the connected socket is closed. This is the point when I want the gets of the sending function to be closed too. It does stop when it fails to send whatever input you give, but that's one useless input I'm trying to get rid of.

Ludovic 'Archivist'

Your thing reminds me of cuckoo hashing, which I did use at some point

Ammar

I didn't provide any breakpoint

Or try gdb -tui ga It shows the code being executed.

Ludovic 'Archivist'

You can find an implementation of this here (albeit not for a hashmap but a quotient filter): https://git.nekoit.xyz/Archivist/gplib/src/branch/master/include/gp/quotient_filter.hpp

Ajay

Or try gdb -tui ga It shows the code being executed.

yeah it shows the correct code

Anonymous

I think memset will do the same.

explicit_bzero time

Anonymous

Reminder that compilers special case functions like memset so sometimes they may not work

Ammar

yeah it shows the correct code

Still no clue, try to press CTRL + C and write x/10i $rip It shows the Assembly instructions being executed.

Ammar

yeah it shows the correct code

It usually shows an arrow to corresponding line, but it does not.

Ajay

yeah it shows the correct code

Actually, I'm using MacOS mojave and so when i sought the problem online, i followed this link which says gdb is not compativle with MacOS sierra(didn't check for mojave)

Ammar

Reminder that compilers special case functions like memset so sometimes they may not work

explicit_bzero is just a prevention for optimization, he needed multibyte fill.

Ajay

Actually, I'm using MacOS mojave and so when i sought the problem online, i followed this link which says gdb is not compativle with MacOS sierra(didn't check for mojave)

https://stackoverflow.com/questions/40110435/unknown-ending-signal-when-using-debugger-gdb

Ajay

aah man, earlier I was doing ctrl+z which was working(stopped the running program) because ctrl+c wasn't but now even ctrl+z isn't working.

Ammar

https://stackoverflow.com/questions/40110435/unknown-ending-signal-when-using-debugger-gdb

I never used gdb on Mac OS. Don't really know if it is not compatible.

Ajay

I never used gdb on Mac OS. Don't really know if it is not compatible.

np

Ammar

Try valgrind sir.

Sasuke

My program uses SDL library which is cross platform but installation is needed for every platform which is different accordingly so I wanted to ship my program without hassle to the user like installing a compiler and installing required libraries how can I do that?

Ajay

Try valgrind sir.

g++ -g filename.cpp valgrind ./a.out is this right?

Ammar

g++ -g filename.cpp valgrind ./a.out is this right?

Yes, some additional flags may be suggested in runtime.

Ajay

Yes, some additional flags may be suggested in runtime.

I just wanted the segmentation fault to be caught, I hope just that to work.

Ammar

I just wanted the segmentation fault to be caught, I hope just that to work.

Try to run a segfault code then

Ajay

Try to run a segfault code then

yes, It is showing Invalid write of size 4 ==6760== at 0x100000DA9: solve() (ga.cpp:11) ==6760== by 0x100000DC2: main (ga.cpp:17)

Ammar

Good, happy to work.

Ajay

But, when I simply create int a[3000000]; inside main and run then why doesn't that give segmentation fault? But, it does give segfault in that code. Any reason?

Ammar

But, when I simply create int a[3000000]; inside main and run then why doesn't that give segmentation fault? But, it does give segfault in that code. Any reason?

Your program may still have available stack. When you start to call other functions your stack will start increasing until the called function reaches epilogue (return).

Ajay

Your program may still have available stack. When you start to call other functions your stack will start increasing until the called function reaches epilogue (return).

I wish we could see the stack space allocated to my current cpp process.

Ajay

I tried experimenting doing long long a[3000000000000000000000]; then the compiler says error: size of array 'a' is negative, why? Even adding ll at the end of that number says that too.

Ammar

I wish we could see the stack space allocated to my current cpp process.

Probably can't, but you can see the max limit stack by using getrlimit.

Ajay

long long is ~ 1e18, so i think that shouldn't be the problem

Ammar

I tried experimenting doing long long a[3000000000000000000000]; then the compiler says error: size of array 'a' is negative, why? Even adding ll at the end of that number says that too.

Integer overflow may cause negative value, try to use ull for unsigned, so it never be negative.

Anonymous

The largest array I've ever seen was like 3000 in size

Ajay

oh, didn't see that 😅