at 13:00 you mention that java ‘already passes by reference’; as the smart way does in c, shown by you as the 3rd way. however i’m missing a finer distinction between reference, pointer and address, e.g. should reference and pointer be equaled, are they working the same? to me structs and arrays in c are addresses.
There is either "pass by reference" or "pass by value". In C... everything is passed by value. Pointers hold addresses but, when you call a function, you actually pass the value of that pointer to that function (the value, in this case, being an address). When passing a struct, the program makes a copy of it on the stack of the function called. When passing an array you pass by value the address to the first element. In higher level languages (which are garbage collected), objects and arrays are passed by reference instead and the compiler does this dereference operation for you. Usually just primitives (int, bool, char etc.) are passed by value because it's cheaper to simply copy the value and send it to the function
So is the malloc only required for things where we dont know the size of at compile time? For example lists of objects? Everything else could be done by using the stack allocation?
Yes, most of what you need can be done on the stack. There's a stack memory limit of 1MB (but it depends). So also, if you need a variable that uses more than 1 MB you would need to use malloc for it
I want to return the pointer of a dynamically allocated struct from a function. But my program is a gui library. The pointer I want ti return is a TreeNode struct for a TreeView. So I want to keep those pointers in a list and when the TreeView get destroyed, I want to check the status of that pointers and want to free them one by one. Is it possible ? This is to avoid the user from the burden of freeing the tree nodes..
A good practice is to leave the resposibility of deallocation to whoever allocated that data. But, if you want to implement it that way you can always assign NULL to pointers that have been deallocated. Although you might need to pass a double pointer to your function. Here is a video related to the topic: code-vault.net/lesson/rh1tys8m13:1603733526917
@@CodeVault thanks a lot. Also i have one more doubt. Inside function if we declare structure using pointers but without malloc(ex: point* m;)where it will store, either stack or data segment?
Any variable you declare in the function is stored on the stack. point* m; Although, this variable is a pointer (an address), not the struct itself. The struct doesn't exist yet. Doing this: m->x = 15; Would throw an error. You create the struct using malloc or you could set the pointer to point to some other point struct: point* m; point n; m = &n; m is a pointer to the struct n here. You have just 1 struct in this piece of code here basically
How come we are able at 6:37, this works? We are trying to access a variable that got popped from the stack, for example if the function returned int*, and we tried to access that array that got allocated in the stack, we would even get a warning because we are trying to return the address of a local variable.
It works only because I deallocated the memory right when it got free. The code essentially looked at some garbage values that just happened to still have the same value as before the deallocation in them. If you wait a while before deallocating you will either get completely random values or a segmentation fault because that part of memory is used by another process
As someone who learns Java it is weird to me how structs are copied in C. It is weird to me that instances of structs aren't pointers. I don't understand what they are if they are not pointers.
Well, with Java you pass around pointers all the time except for primitives. So, when passing around a struct in C it's similar to passing around a primitive (like an int) in Java. You basically pass a **copy** of that struct and not just a reference to it.
If we check the size of the struct example: struct {char a;int b; double c;}; how the structure padding works here? Here size is 16 bytes. if 32-bit processor then 1word=4bytes, if 64-bit processor 1word=8bytes, let us consider i am working with 32-bi processor char =1bytes+ 3bytes extra for padding , int =4 bytes and double = 8bytes(4 + 4) then, size of this struct is 16bytes. suppose in this case struct{char a; int b; double c; char d;}; in this scenario how size will allocate here?
The video is quite outdated and partially wrong. But basically the padding makes sure each member of the struct is at an address multiple of its size. So: a would be 1 byte at address 0 of that struct b would be 4 bytes at address 4 of that struct c would be 8 bytes at address 8 of that struct d would be 1 byte at address 16 of that struct The struct itself will be of size 24 bytes on 64-bit arch and 20 bytes on 32-bit arch Here's a program that you can run to test: #include #include typedef struct t{ char a; int b; double c; char d; } t; int main(int argc, char* argv[]) { printf("%zu ", sizeof(t)); printf("%u %u %u %u ", offsetof(t, a), offsetof(t, b), offsetof(t, c),> return 0; } For 64 bit compile: gcc -o main main.c For 32 bit compile: gcc -o main main.c -m32
Very well explained, better than I tend to do 😅 In my old job we had interns that I always had to teach to stop returning a value like it is done in Pascal or Java. And I was very confused to find out that in college they didn’t learn that arguments go on the stack. That a stack pointer just moved back up when you exit and indeed if you had another thread running the stack could’ve been changed. We had that happen. And ironically those Java, Python and pascal people hate seeing a pointer being passed to home a result. That’s a foreign concept to them. And that’s why I am a firm proponent of at least having students code in assembly for the first semester. Even on a simple 8 bit micro doesn’t need to be very complex. Just to really understand how memory is used.
@@CodeVault I agree, to me the knowledge of assembly really helped me understand C. It doesn’t matter what cpu you learned assembly on. The fundamentals are the same from 6502, z80, 68000 to arm and x64. It also really helps me in debugging core dumps. Unfortunately it’s getting rare that students get assembly here in The Netherland at least.
Strange. I learned python first semester, and only on the third I'm learning C and system design, followed by assembly on the fourth semester. This is dumb that new programmers aren't given the opportunity to learn memory in depth.
@@IvanScriptsEverything I don’t understand the order either. I do understand that they start with a higher language just to get comfortable and not immediately scare everybody away. But then you’ll need to so assembly and then C. You’ll appreciate C and that higher language all the more. We did pascal 1 semester and then assembly and the rest of college it was C.
I really hope you get more subscribers man, your video are by far the best when it comes to c programming.
Mastering C feels like learning some sort of arcane magic...
At least it's better than learning Assembly
@@whenthethebeansstrikeback6728 The forbidden scrolls (punch cards..?)
@@batugunduz3950 Vacumn Tubes
Only when you’ve never developed in assembly before. Otherwise dit becomes like a blessing m 😊
Finally understand why so many library functions ask for a pointer for result instead of using a return value.
It took me a while, but this has got to be the best video on this topic!
What a fantastic lecturer! Always straight to the point (and to the subtlest and most important ones!). Your explanations are always cleanly compiled.
Thank you! I try my best. These types of comments motivate me to keep going
This is a really worthwhile and best video on returning the structs!
Thank you so much for this, you are truly great!
This was such a great video, I really learned a lot. Thank you so much!
Thanks bro.... you just cleared my confusion about C++ and C# about the default function calling type
Excellent and on point explanation. Thank you.
Thank you so much for this video
Thank you! Super helpful.
thank you man
Keep up the good work!!
CAN WE GET A TUTORIAL ON BUILDING A PROJECT WITH C... SEEMS I HAVE DONE ENOUGH ON THE LEARNING ASPECT LEFT WITH THE APPLICATION
That was simple amazing. Step by step, explaining the thought process at each one of them, until reaching the perfect code... Thanks!!
small struct like point and any kind of pointer wrapper I would pass them by value
Hi, Nice vídeo, Very easy to understand, thank you.
Fantastic videos. Your instruction is very good.
Great video mate! Rewards from Argentina.
as you say (AMAZING). thank you
I changed my mind you are actually quite cute
clear explanations, thanks!
at 13:00 you mention that java ‘already passes by reference’; as the smart way does in c, shown by you as the 3rd way. however i’m missing a finer distinction between reference, pointer and address, e.g. should reference and pointer be equaled, are they working the same? to me structs and arrays in c are addresses.
There is either "pass by reference" or "pass by value". In C... everything is passed by value. Pointers hold addresses but, when you call a function, you actually pass the value of that pointer to that function (the value, in this case, being an address). When passing a struct, the program makes a copy of it on the stack of the function called. When passing an array you pass by value the address to the first element.
In higher level languages (which are garbage collected), objects and arrays are passed by reference instead and the compiler does this dereference operation for you. Usually just primitives (int, bool, char etc.) are passed by value because it's cheaper to simply copy the value and send it to the function
So is the malloc only required for things where we dont know the size of at compile time? For example lists of objects? Everything else could be done by using the stack allocation?
Yes, most of what you need can be done on the stack. There's a stack memory limit of 1MB (but it depends). So also, if you need a variable that uses more than 1 MB you would need to use malloc for it
thank you, helped a lot! :D
I am not able to understand some of your videos. What should i do. I think they are some high level.
Exercise on your own. Take the code I am explaining and try to change it little by little. Also, research the documentation thoroughly
I'm from the 80's and your video is really awesome how you expose through levels the right use of pointers in a "healthly" way. Thanks for sharing!!.
I want to return the pointer of a dynamically allocated struct from a function. But my program is a gui library. The pointer I want ti return is a TreeNode struct for a TreeView. So I want to keep those pointers in a list and when the TreeView get destroyed, I want to check the status of that pointers and want to free them one by one. Is it possible ? This is to avoid the user from the burden of freeing the tree nodes..
A good practice is to leave the resposibility of deallocation to whoever allocated that data. But, if you want to implement it that way you can always assign NULL to pointers that have been deallocated. Although you might need to pass a double pointer to your function. Here is a video related to the topic: code-vault.net/lesson/rh1tys8m13:1603733526917
@@CodeVault Thank you very much for your suggestion. I didn't think about the double pointer.
bcuz of your videos i am starting to understand and like C. Thank you.
Thanks ❤
bravo!
the fact that he doesnt look at notes or script shows me how well he knows his stuff in contrast to others
You are such a good instructor! Thanks!
very good vid helped quite a lot
the ---> Point* getmiddlepoint has point m struct inside so struct inside struct... right???
getMiddlePoint is just a function and m is a local variable for that function
Awesome! What about declaring a "static Point" at getMiddlePoint function, and return the pointer?
That would work as well
I get the error incompatible types when assigning to type 'Point (aka struct Point)' from type 'int'
Hmm, what's the code you have? You're assigning something wrong to the Point variable
if we return structure, what exactly it will return. which means it return structure values or structure address?
Structure values. It basically returns a copy of the structure (same as how you return a primitive like an integer)
@@CodeVault thanks a lot. Also i have one more doubt. Inside function if we declare structure using pointers but without malloc(ex: point* m;)where it will store, either stack or data segment?
Any variable you declare in the function is stored on the stack.
point* m;
Although, this variable is a pointer (an address), not the struct itself. The struct doesn't exist yet. Doing this:
m->x = 15;
Would throw an error. You create the struct using malloc or you could set the pointer to point to some other point struct:
point* m;
point n;
m = &n;
m is a pointer to the struct n here. You have just 1 struct in this piece of code here basically
@@CodeVault Thanks a lot @CodeVault
Thank you bro, very clear!! where are you from??? i'm Colombian, and way older than you!!
I'm from Romania
How come we are able at 6:37, this works? We are trying to access a variable that got popped from the stack, for example if the function returned int*, and we tried to access that array that got allocated in the stack, we would even get a warning because we are trying to return the address of a local variable.
It works only because I deallocated the memory right when it got free. The code essentially looked at some garbage values that just happened to still have the same value as before the deallocation in them. If you wait a while before deallocating you will either get completely random values or a segmentation fault because that part of memory is used by another process
As someone who learns Java it is weird to me how structs are copied in C. It is weird to me that instances of structs aren't pointers. I don't understand what they are if they are not pointers.
Well, with Java you pass around pointers all the time except for primitives. So, when passing around a struct in C it's similar to passing around a primitive (like an int) in Java. You basically pass a **copy** of that struct and not just a reference to it.
If we check the size of the struct example: struct {char a;int b; double c;}; how the structure padding works here? Here size is 16 bytes. if 32-bit processor then 1word=4bytes, if 64-bit processor 1word=8bytes, let us consider i am working with 32-bi processor char =1bytes+ 3bytes extra for padding , int =4 bytes and double = 8bytes(4 + 4) then, size of this struct is 16bytes. suppose in this case struct{char a; int b; double c; char d;}; in this scenario how size will allocate here?
The video is quite outdated and partially wrong. But basically the padding makes sure each member of the struct is at an address multiple of its size. So:
a would be 1 byte at address 0 of that struct
b would be 4 bytes at address 4 of that struct
c would be 8 bytes at address 8 of that struct
d would be 1 byte at address 16 of that struct
The struct itself will be of size 24 bytes on 64-bit arch and 20 bytes on 32-bit arch
Here's a program that you can run to test:
#include
#include
typedef struct t{
char a;
int b;
double c;
char d;
} t;
int main(int argc, char* argv[]) {
printf("%zu
", sizeof(t));
printf("%u %u %u %u
", offsetof(t, a), offsetof(t, b), offsetof(t, c),>
return 0;
}
For 64 bit compile: gcc -o main main.c
For 32 bit compile: gcc -o main main.c -m32
@@CodeVault thanks a lot
thank you so much
Very well explained, better than I tend to do 😅 In my old job we had interns that I always had to teach to stop returning a value like it is done in Pascal or Java.
And I was very confused to find out that in college they didn’t learn that arguments go on the stack. That a stack pointer just moved back up when you exit and indeed if you had another thread running the stack could’ve been changed. We had that happen.
And ironically those Java, Python and pascal people hate seeing a pointer being passed to home a result. That’s a foreign concept to them. And that’s why I am a firm proponent of at least having students code in assembly for the first semester. Even on a simple 8 bit micro doesn’t need to be very complex.
Just to really understand how memory is used.
Yeah, in my University we learned assembly in the first semester. It was incredibly helpful in understanding the rest of the languages
@@CodeVault I agree, to me the knowledge of assembly really helped me understand C. It doesn’t matter what cpu you learned assembly on. The fundamentals are the same from 6502, z80, 68000 to arm and x64. It also really helps me in debugging core dumps.
Unfortunately it’s getting rare that students get assembly here in The Netherland at least.
Strange. I learned python first semester, and only on the third I'm learning C and system design, followed by assembly on the fourth semester. This is dumb that new programmers aren't given the opportunity to learn memory in depth.
@@IvanScriptsEverything I don’t understand the order either. I do understand that they start with a higher language just to get comfortable and not immediately scare everybody away. But then you’ll need to so assembly and then C. You’ll appreciate C and that higher language all the more.
We did pascal 1 semester and then assembly and the rest of college it was C.
How about passing and returning an array of structs?
In this video I show how you can return an array: code-vault.net/lesson/iyvrv54u8e:1603733520962
Great job!