Simple array issue

[sinful hamlet]

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 run !howto ask.

[tulip anchor]

you can't use a pointer to initialize an array like that. Since VLAs are bad, your best bet will be to either allocate new memory of correct size and copy over the contents of y_coords, or just use y_coords directly

[open vale]

How would I do that? Sorry if this is a stupid question, but I'm trying to translate from Python

[tulip anchor]

you could use malloc to allocate new memory (don't forget to free it when you're done)

though y_coords can just be used directly

[open vale]

Unfortunately, this won't work for what I'm trying to do.

[tulip anchor]

then the allocate-and-copy path seems like the move

[open vale]

You are suggesting I use malloc and then copy each element?

[tulip anchor]

no, allocate enough memory for the entire array, then copy over the full contents of y_coords in a single move

[open vale]

double d2_coords[] = malloc(Nj * sizeof(double));
```How would I then set it to `y_coords`?

@tulip anchor

[tulip anchor]

you'd need a double*, not a double[], for d2_coords

and you can copy it over with memcpy

[open vale]

Would that work with my use case?

[tulip anchor]

again, the dn_coords need to be double*, not double[]

and then you're still trying to use = to move the data over, which won't work. use memcpy

[autumn oar]

that's a lot of arguments for a function

[tulip anchor]

also that

you're also re-declaring dn_coords right after mallocing them, which won't compile

[open vale]

What would be the best way to solve that?

[tulip anchor]

you probably want Nx * sizeof (double) as the size for the memcpy calls

now you have a double*[]. you just want a double*

[open vale]

FINALLY. Thank you I really appreciate it.

[tulip anchor]

I need to see the code that's actually generating the error

[open vale]

Split in two because of max message length

[sinful hamlet]

@open vale

Please Do Not Delete Posts!

Please don't delete forum posts. They can be helpful to refer to later and other members can learn from them. In the future you can use !solved to close a post and mark a post as solved.

[tulip anchor]

values3D is a double* but you're trying to subscript it twice

[open vale]

So should I use the & or how does that work

Again, sorry about the really basic questions

[tulip anchor]

you should change the parameter type to actually match what you're trying to pass into it and use it as

looks like you're passing this into it:

double rho[Ni][Nj][Nk];

[open vale]

Right

[tulip anchor]

so besides the whole issue with VLAs being bad, that is not a double*

[open vale]

What is a VLA?

[sinful hamlet]

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.
VLAs have poor compiler support and can lead to inefficient code. The core issue with VLAs is that the compiler doesn't know the size of the stack frame. Without warning flags like -Wvla (Note: -Wvla is not turned on by -Wall nor any other warning flag) it can be easy to create a VLA by accident, even in C++ with some compilers.

Compiler Support

✅ available since C99 ⛔ not available in C++ at all ⛔ was never supported by MSVC ⚠ optional feature since C11 ⚠ supported as non-standard extension by GCC, clang

See Also

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

[open vale]

I thought it's not a VLA because I use const?

[open vale]

I'm not really sure where this is a problem?

[tulip anchor]

that isn't what actually does it, but you're right. it's not a VLA

I didn't check where Ni and such came from

[autumn oar]

const means cannot change, not that it is known at compile time

const int x = foo();
int a[x]; // the size can vary

[open vale]

Where is double* not used correctly?

[tulip anchor]

in your function signature

you have it as just a double* which is what a 1D array of double would decay into

your 3d array would decay into double (*values3d)[Nj][Nk]

[open vale]

inter_d2_values[d2] = inter_1D(d1_coords, double (*values3D)[d3][d2], Nd1, pos_d1);

Like so? I'm confused about what we are talking about

[tulip anchor]

in your 3D function, not your 1D function

and in the signature, not the call

[open vale]

How would I change my function call afterwards?

[tulip anchor]

look at the error. You're declaring values3D in the parameter list as just a double*, so the first layer of subscripting will get you a double. You're then trying to subscript that single double again, which isn't going to work. If you instead declare values3D to be a double (*)[Nj][Nk], (a pointer to double[Nj][Nk]) then the first subscript will get you a double[Nj][Nk], and then the second subscript will get you a double[Nk], which will decay into the double* that inter_1D expects

[open vale]

That makes sense to me

[sinful hamlet]

@open vale Has your question been resolved? If so, run !solved :)

[open vale]

!solved