Nice work! Added a standard compile chain to have a look but doesn't compile "as is". Feel free to borrow.

Dockerfile: builder.Dockerfile

ARG base_tag=bullseye
ARG base_img=mcr.microsoft.com/vscode/devcontainers/base:dev-${base_tag}

FROM --platform=linux/amd64 ${base_img} AS builder-install

RUN apt-get update --fix-missing && apt-get -y upgrade
RUN apt-get install -y --no-install-recommends \
    apt-utils \
    curl \
    cmake \
    build-essential \
    gcc \
    gdb \
    g++-multilib \
    locales \
    make \
    ruby \
    gcovr \
    wget \
    python3 \
    && rm -rf /var/lib/apt/lists/*

Makefile:

# project 
project_name=c_std

builder-build :
	docker build \
                    -f builder.Dockerfile 
                    -t ${project_name}-builder:latest .

builder-run :
	docker run \
		--rm \
		-it \
		--platform linux/amd64 \
		--workdir /builder/mnt \
		-v ${PWD}:/builder/mnt \
		${project_name}-builder:latest \
		/bin/bash
                #/bin/bash -c "python3 compile.py b"

Compile:

make builder-build
make builder-run

(container) python3 compile.py b

Output:

...truncated...
file_io/file_writer.c: In function ‘file_writer_lock’:
file_io/file_writer.c:494:15: warning: implicit declaration of function ‘fileno’ [-Wimplicit-function-declaration]
  494 |     if (fcntl(fileno(writer->file_writer), F_SETLKW, &fl) == -1) {
      |               ^~~~~~
file_io/file_writer.c: In function ‘file_writer_truncate’:
file_io/file_writer.c:580:9: warning: implicit declaration of function ‘ftruncate’; did you mean ‘strncat’? [-Wimplicit-function-declaration]
  580 |     if (ftruncate(fd, size) == -1) {
      |         ^~~~~~~~~
      |         strncat
Compilation failed.

1. Excess duplicate parameters
   

2. Not including necessary header file path
   

//json.c:621
    if (state.current_token.type == JSON_TOKEN_OBJECT_START) {
        root = parse_object(&state);

    } 
    else if (state.current_token.type == JSON_TOKEN_ARRAY_START) {
        root = parse_array(&state);
    } 
    else {
        fmt_fprintf(stderr, "Error: Root element must be an object or array at position %zu.\n", state.position);
        free(state.input); 
        return NULL;
    }

But a json can be valid as a single and unique value

"value"

53

etc...

I noticed there is no Trie (prefix tree) module yet. I am almost finished writing one then I'll initiate a PR. Apologize for any time lag in advance, I work on it when day job isn't occupying all my time.

I may add all issues about vector_push_back,vector_emplace_back and related for reference to check

// json.c:789
        case JSON_ARRAY:
            for (size_t i = 0; i < vector_size(element->value.array_val); ++i) {
                JsonElement *item = json_deep_copy(*(JsonElement **)vector_at(element->value.array_val, i));
                if (!item) {
                    json_deallocate(copy);
                    return NULL;
                }
                vector_push_back(copy->value.array_val, &item);
            }
            break;

What is expected here for example if vector_push_back fails ? Should not we free all the copied data and return NULL ?
Because user may have a error message but nothing is recoverable from here...

Platform Ubuntu22.04

Hi, when I test your code I find a good way to solve it without use the FMT_END_ARGS in fmt.c locate in fmt_print and fmt_println.

Recurrence

when i use the api , i don't want to use it with the define FMT_END_ARGS just like the following code.

int main() {
    const char* str = "Hello, World!";
    fmt_println(str);
    return 0;
}

the output is

Compilation successful. Output: ./build/main
Running the program...
*************************
Hello, World! �L�� �L�� @!�I�U  ��UH��AWAVAUATSH��8L�%�I 

How to Fix it

There is a high probability that garbled characters appear because you cannot use simple define va_arg() to show all parameter list in C code. It causes undefined behavior.

The description of this project is clickbaity and misleading.

This is NOT the C++ Standard Libraries, it's a try-implementation of the C++ std libs.

Vulnerability-fixes and Performance is not even close to the standard c++ implementation.

Please change the description to better address this, you can change to the similarities of "Basic C++ libraries for C", just don't put "THE STANDARD".

Do you mind considering using NULL pointer or -1 as return value for when there is an error instead of exit(-1) ? It isn't safe to use with exit(-1) because most of the time we don't want an application to crash because of a simple error even if it's memory related.

But also, is it possible to guard each call to fmt_fprintf with a "#ifdef LOGGING_ENABLE" or "#ifndef LOGGING_DISABLE" ? We don't want all the time printing to stderr libraries error in client applications

    if (vec->size == vec->capacitySize) {
        vector_reserve(vec, vec->capacitySize * 2); // Use the modified version
    }
    char *base = (char *)vec->items;
    memmove(base + (pos + 1) * vec->itemSize, 
            base + pos * vec->itemSize, 
            (vec->size - pos) * vec->itemSize);
    memcpy(base + pos * vec->itemSize, item, vec->itemSize);
    vec->size++;

What is the vector supposed to do when reserve fails this is actually a really dangerous matter as we may overwrite data. There shall be a way to inform the user of problems (other than just printing out a message)

Simply returning a boolean maybe ?