Should I try to completely understand everything about C before moving on to newer languages ike C++

[hidden tusk]

Dumb question I know but I just wanted to ask how much of C should I try to understand before moving on to other languages since my professor Is done with C and will most likely move on to C++ so I wanted to be done with C first since there is a lot of stuff I still don't know about it

[steep cedar]

When your question is answered use !solved to mark the question as resolved.

Remember to ask specific questions, provide necessary details, and reduce your question to its simplest form. For tips on how to ask a good question use !howto ask.

[chilly orchid]

there is a lot of stuff I still don't know about it

it's fine

[hidden tusk]

It is?

I don't know every time I ask a question here I find out about 10 more things

[chilly orchid]

when I passed my linear algebra lesson for the first time, I didn't rank quite well (a poor 14/20). the affine geometry lesson next semester, was entirely based on that one class. I studied for that exam, and presented that one and the linear algebra one again, to increase my score (without studying, because the teacher was nice: he kept the best score for students that represented in June, as a favour. since I had many other exams, I just didn't study and tried anyway). I got 18 and 19, respectively.

[chilly orchid]

you might review C concepts while learning C++, get haha-moments, and understand better anyway

there are different theories, but often people tend ot agree that good programmers are programmers that know different languages on surface (because it shows they came accoss different paradigms and ideas), and are able to dig deep into one when required (because they understood the concepts). knowing all C by heart, doesn't sound like a requirement right now

[hidden tusk]

Hmm...Oke I'll stick with C only for a bit just to really understand what I learned and then Move on to C++ since there is a big gap before we start it anyways

[hidden tusk]

Oh I think I at least understand the concepts of it

[chilly orchid]

that's already nice

[hidden tusk]

Oke then I think I'll be moving on to C++ after a bit then

Thanks a lot for the help man

[bright mulch]

I hope sincerely that you are eased into C++ and not dropped straight into the deep end. Modern C++ scares even me with all the templating and whatnot.

[chilly orchid]

don't forget to come back! 😉 life's better here
(it's a joke 😄 well, I mean it, but do what you prefer 😄 hahaha)

[bright mulch]

I know that there will be things you will immediately love and you may even start criticizing C somewhat.
A very simple example of that being function overloading.

[hidden tusk]

Yee my prof is good since it was my first time programming and I Aced my mids due to his help...just hope it continues on to C++

[hidden tusk]

Yee I just hope it's not that hard😭

[hidden tusk]

Oh wait is that where you couldn't have 2 main functions of same name?

[neat hemlock]

I would say it is a good idea to learn C++ and since you already know some C it won't be as difficult as a newcomer to the languages.

[bright mulch]

void foo(int arg);
void foo(float arg);

int main()
{
  int i = 5;
  float f = 5.0;
  foo(i);      // calls the int version
  foo(f);      // calls the float version
}

[neat hemlock]

functions of the same name with different parameters

also I don't think you can overload the main function

[bright mulch]

No, even in C++ you cannot overload main().

[hidden tusk]

Oh

So it's just for multiple user defined ones with same name?

[neat hemlock]

templates are one of the main things that make me want to stay writing C++ over C

[bright mulch]

Going from naught to C++23 will scare most.
I hope that you start off aiming for C++03 or perhaps C++11.

[neat hemlock]

The STL is also amazing compared to not having it at all

[neat hemlock]

I would say C++14 since it kind of "fixes" some of the missing things in 11

But yeah totally agree C++23 is massive

[hidden tusk]

Wait it has numbers? I'm Assuming you start from 01

[bright mulch]

Those numbers are the years that the then new standard came out.
C++03 came out in 2003.

[hidden tusk]

Ohhh

[bright mulch]

This site has been most valuable to me. It may help you.
https://github.com/AnthonyCalandra/modern-cpp-features

[hidden tusk]

Damn 23? It's still getting updates?

[neat hemlock]

Just for reference, this site will be very helpful https://en.cppreference.com/w/

Not a tutorial site though

[neat hemlock]

C is also still getting updates

There's C23 (or C2X I don't remember)

[bright mulch]

cppref has already started mentioning C++26 I noticed the other day.

[hidden tusk]

I'll have it saved thanks alot

[lost glen]

is there a calendar for C features?

[bright mulch]

Like a roadmap?

[lost glen]

yeah, as it seems to be a roadmap

[neat hemlock]

@bright mulch Have you used coroutines in any code yet?

[hidden tusk]

Are they significant? I thought most of the functions we used were from C99

[neat hemlock]

I personally haven't - but I really never see a good time to utilize it (and also am still new/learning it)

[neat hemlock]

not really

the site I mentioned also has information on C too

[bright mulch]

Nope.
I work mostly in embedded these days.
Those compilers fall way, way back on standards. Always have.
Maybe if one used the GNU compilers, then you could use modern C++.

As it is, I am (still) tied to C++17.

[hidden tusk]

Yee I see some are mentioned (C11)

[neat hemlock]

Hey that's not too bad! At least you're not on C++11 - not saying I am but that would not be great

[neat hemlock]

Yeah there's also a link for C23

[hidden tusk]

There is also a C26 wat does that mean?

[neat hemlock]

Wait where

Are you talking about C++26?

[bright mulch]

If true, it seems that 2026 is a unique year in that it brings out new standards for both C and C++.

[hidden tusk]

Yes

It says on that website

[neat hemlock]

I think they are supporting reflection in C++26

[hidden tusk]

Mb I accidentally said C26 instead

[neat hemlock]

if so that should be interesting

[hidden tusk]

Oh

Damn

[bright mulch]

To be honest, even though C++26 has been announced, its contents will be quite fluid.
Some stuff may not be ratified.
Other stuff may be brought into the mix.
A roadmap for C++26 would be largely a work of fiction.

[hidden tusk]

Man I can't wait to just jump into C++ now after my finals

[neat hemlock]

Yeah, should be a lot of fun

[neat hemlock]

For sure, a lot of what's listed is definitely not going to be delivered

[hidden tusk]

Yee I hope so

[polar lintel]

im thinking

[neat hemlock]

Also like the idea of std::hive

[polar lintel]

i went into C++ before i knew the basics of C and genuinely

[neat hemlock]

That one is intriguing

[polar lintel]

i still think C is harder 💀

i literally went down the ladder instead of up (C#->C++->C) and ngl C just makes it way harder to do anythin

[neat hemlock]

yeah there are less abstractions which make it seem harder

[polar lintel]

ie no inheritance without some kind of fuckery

[neat hemlock]

C strings are pretty sad compared to C++ strings

[polar lintel]

on god

i still dont know how to read anything of any size from a stream

is that even possible

[neat hemlock]

string?

strlen

[polar lintel]

with a C string

[lost glen]

CStrings and CPPStrings are different?

[polar lintel]

fgets(buf, 1024, stdin);
``` in example i dont know how to make that 1024 entirely variable

[neat hemlock]

yeah char * cstr;
std::string cppstr;

[lost glen]

I think you mean string utils, If not, I suppose cppstrings keep the null terminator rule

[neat hemlock]

Yeah

[bright mulch]

But, with C you are in complete control, so this is why it is so popular with embedded.
For instance, using the STL you think arrays (is: vector) just became 100x easier.
But if you let it, STL will use heap for anything dynamic.

That's the stuff of nightmares in embedded.
So then you need to learn allocators to ensure that you control dynamic allocations.
And suddenly something as simple as arrays isn't so simple after all.

[neat hemlock]

I had to do this in a pacman project, but I don't remember the details

[polar lintel]

how do you do dynamic arryas without heap

[bright mulch]

allocators.

[lost glen]

you can allocate scalable memory within the stack?

[polar lintel]

also isnt fam and vla technically just heap under the hood too or am i stupid

[neat hemlock]

no idea, I stay away from VLAs

[polar lintel]

same

[lost glen]

why?

[bright mulch]

Example:
In threadx, you can create byte pools.
You can do new and placement new, and redirect from heap to these byte pools.

These byte pools could be a stack buffer.
Or an area in the .bss.

[steep cedar]

What Is a VLA, and Why Is It "Bad"?

A Variable Length Array (VLA) is an array where the size is not constant and depends on a variable.

Example

int size = rand();
int vla[size]; // VLA of type int[size]
int not_vla[10]; // regular array of type int[10]
constexpr int size = 10;
int arr[size]; // also not a VLA, of type int[10]

Why Are VLAs "Bad"?

VLAs have poor compiler support and can lead to unsafe code. The core issue with VLAs is that the compiler doesn't know the size of the stack frame. Without warning flags like -Wvla, it can be easy to create a VLA by accident, even in C++ with some compilers.

Compiler Support

:white_check_mark: available since C99
:no_entry: not available in C++ at all
:no_entry: was never supported by MSVC
:warning: optional feature since C11
:warning: supported as non-standard extension by GCC, clang

See Also

What technical disadvantages do C99-style VLAs have?
What's the point of VLA anyway?

[chilly orchid]

VLA to allocate memory is a really bad idea

[polar lintel]

;compile -Wvla -std=c23

#include <stdio.h>
#include <time.h>
#include <stdlib.h>

constexpr int sqr(int x) {
  return x * x;
}

int main(void) {
  srand(time(NULL));

  int i = rand() % 30;
  int arr[sqr(i)];

  printf("%zu", sizeof arr / sizeof i);
}```

[fleet skiff]

Compiler Output

<source>:5:1: error: 'constexpr' requires an initialized data declaration
    5 | constexpr int sqr(int x) {
      | ^~~~~~~~~
<source>: In function 'main':
<source>:13:11: error: implicit declaration of function 'sqr' [-Wimplicit-function-declaration]
   13 |   int arr[sqr(i)];
      |           ^~~
<source>:13:3: warning: ISO C90 forbids variable length array 'arr' [-Wvla]
   13 |   int arr[sqr(i)];
      |   ^~~
Build failed

kirberburgy | c | x86-64 gcc 14.1 | godbolt.org

[chilly orchid]

VLA to implify size computation, is a good idea

[lost glen]

ah, so It is generally a bad idea to define an array's width with a non-constant integer expression

[polar lintel]

so tha'ts not vla?

[chilly orchid]

int x, y; // assume given at runtime

double m[x][y]; // bad vla usage
double (*m)[x][y] = malloc(sizeof(*m)); // friendly vla usage

[lost glen]

shouldn't be as sqr returns a constant signed integer expression

[neat hemlock]

Also do you have to cast malloc there or is it not necessary

[polar lintel]

why would you need to cast malloc

[bright mulch]

You wouldn't even be doing malloc() in C++.

[neat hemlock]

Since it of type void*

[chilly orchid]

no it's fine

[neat hemlock]

ok

[chilly orchid]

ah wait, we are in C++ ?

[bright mulch]

No, but the question is about C++ after C.

[polar lintel]

new

[chilly orchid]

aaaaah

[bright mulch]

Still a good point (vla), though.

Plus OP would have learned about new vs malloc pretty damn early on anyway.

[neat hemlock]

Also I believe it's necessary to cast there in C++

[polar lintel]

nobody mallocs in c++

[neat hemlock]

but yeah, who needs malloc

[steep cedar]

@hidden tusk Has your question been resolved? If so, type !solved :)

[hidden tusk]

!solved

[steep cedar]

Thank you and let us know if you have any more questions!

This thread is now set to auto-hide after an hour of inactivity