#include
struct Foo {
char a;
float b;
char c;
};
struct Bar {
float b;
char a;
char c;
};
int main()
{
std::cout
The absolute state of C++
Thomas Baker
Other urls found in this thread:
catb.org
youtube.com
twitter.com
Kayden White
>what is struct padding
James Price
Is this not common knowledge? Anyways GCC has an attribute to reorder elements for the smallest size: __attribute__((packed)) afaik
Luke Nguyen
This is to show what exactly? Nothing.
Watch OP being the little skanky bitch that he is runs away and closes his browser window so he doesn't have to get the ass fucking he's about to receive from the people about to respond to this thread.
Grayson Gonzalez
>best solution is to use a compiler-specific extension
Caleb Ross
>This is to show what exactly?
It's to show that the C++ standard can't into compile time optimizations.
Jonathan Reyes
You shouldn't be packing structures unless you have a good reason to do so.
Adam Rogers
Jaxon Baker
>best solution
I didn't imply that. The solution is to rearrange them to have biggest-->lowest
Adam Ward
>you shouldn't be using less memory with zero downsides unless you have a good reason to
u wot m8?
Parker Carter
>not using [repr(packed)]
Nathaniel Gomez
>The solution is to rearrange them to have biggest-->lowest
Which is absolutely fucking retarded. The C++ standard should specify that the compiler needs to do that. There is zero reason for this not to be done automatically. Stop trying to justify the faults of a shitty programming language.
Dylan Gray
Sometimes you really want a different ordering.
Zachary Gutierrez
C/C++ assumes the programmer is smarter than the language and not the other way around, now go back to webdev or python faggot
Ethan Taylor
On most computer architectures it's slower to access memory if it's not bounded to a word. On some computer architectures it's mandatory
William Anderson
If your program relies on the order of the member variables in memory, you are doing something horribly wrong and should not ever code anything.
Nathaniel Bennett
>this guy has no idea what padding is and why it was invented
Luke Perry
>moving goalposts
False, which is why one of the greatest C++ lectures ever is about how programmers often shoot themselves in the foot by trying to help the compiler: youtube.com
James Allen
No, spouting absolute utter bullcrap like that and pretending you know better means you are a dangerous idiot. You should fuck off and never ever code anything
Isaac Morales
>webdev baby
Matthew Wright
I know exactly what it is. However, there is literally zero justification for the compiler not optimizing the order of the member variables. Are you really this stupid that you think it's a good thing that the programmer has to do this work manually even though the compiler can easily figure out the most memory efficient order deterministicly?
Colton Watson
Based
Nicholas Barnes
I notice you haven't given any examples of code which relies on the order of member variables and isn't total garbage. The reason why is because such a thing does not exist. If you think it does, you really have no business on this board.
Asher Bell
>sometimes you want your program to consume more memory without any practical advantage
my sides are in orbit
Owen Roberts
If you really need the order to be something specific, you should be using tuples. Even then, you're using them wrong.
Levi Taylor
>literally makes ad hominem attack due to an inability to properly defend his position.
Cool story.
Brandon Brown
there are common cases where you want to have explicit order, wonder if it plays some role in (de)serialization implementations repying on the same layout across machines and how (if it exists) rust deals with it
any paper exploring a different alignment effects on performance?
Aiden Wood
That only relies on the order being consistent, but not on them being in any specific order. All that has to happen is the spec needs to specify what order the compiler needs to put them in, in a way which is not open to interpretation.
Ethan Ward
Still more efficient than you shitlang pajeet
Samuel Hill
it's crucial for cache optimization as well, both automatic and manual.
Levi Taylor
>Is this not common knowledge?
To anyone who studied electronics or compiler design, it is. To stupid Java "developers," it isn't.
Thomas Campbell
the world would probably be a better place if the compiler automatically optimized the order, yes. however, now that we have lots of C-like APIs where you pass around void pointers and where the API relies on the order of the struct data, we can't have that optimization or it will break legacy code.
Eli Watson
C and C++ compilers are not able to do this by standard, which mandates that all struct fields be in increasing address order. If you think this is retarded, then you have never parsed a binary file or network packet in your life.
Rather than bitching that the compiler doesn't perform optimizations that can result in incorrect behavior, you should write your structs the right way the first time around so that they are laid out in memory as you desire it. This is a systems language. You are expected to think about how data fits in memory when you write software.
Asher Robinson
This happens because struct alignment. To avoid this, use pragma pack(push, 1)