remove the use of implicits that are not warp-memory

[AlejandroLabourdette]

I'd suggest you to add warp_memory implicit rather than completely getting rid of it

#[implicit(warp_memory)]
fn f(...) {
  ...
  warp_memory.read_u128(index);
  ...
}

It would be difficult later to track places where warp_memory needs to be added if we remove it from all of the places now.

I agree 100% with you, actually that's was what I think, that's why on CairoFunctionDefinition implementation I appended a comment of how we plan to use implicit and sort of what's happening with them. https://github.com/NethermindEth/warp/blob/5cc3efe06dfff437b11b68929a0385b557bbbfba/src/ast/cairoNodes/cairoFunctionDefinition.ts#L18

Right now we still have a set of Implicits foreach function definition. Even considering that the only available implicit is warplib_memory the logic we had to infer the implicits of the current function worth it.

Nevertheless I do remove them from cairoUtilFuncGen scripts that were not updated to Cairo1 syntax, mainly because we had them hardcode in there (was just a string that we insert in the raw string definition of those functions). They didn't have any role at ast level and we do have to review each of of those script to make them Cairo1 compatible. This is an example: https://github.com/NethermindEth/warp/blob/db1b5987c7fe19c0f7b26dc80239f819be7f644e/src/cairoUtilFuncGen/memory/implicitConversion.ts#L51 https://github.com/NethermindEth/warp/blob/db1b5987c7fe19c0f7b26dc80239f819be7f644e/src/cairoUtilFuncGen/memory/implicitConversion.ts#L210

[piwonskp]

I'd suggest you to add warp_memory implicit rather than completely getting rid of it

#[implicit(warp_memory)]
fn f(...) {
  ...
  warp_memory.read_u128(index);
  ...
}

It would be difficult later to track places where warp_memory needs to be added if we remove it from all of the places now.

I agree. Let's just put #[implicit(warp_memory)] in front of those functions, but without updating the rest of their implementation and syntax

[rjnrohit]

Oops ! conflicts has come up since I have merged the uintN pr, could you resolve it?

[AlejandroLabourdette]

As part of this PR you should update cairoParsing.ts and make it parse the custom implicit attribute in order to detect when generated functions requires the warp_memory implicit

Actually that's already done, we just have the remains of the old parseImplicit, but using getRawCairoFunctionInfo is already known the implicit used (if used any). I'm gonna remove that old function right now

[piwonskp]

Does parseMultipleRawCairoFunctions work properly?

[piwonskp]

There is also getImplicitsAt in ast.ts. Is it needed?