fz-lang/src/compiler/fzl_reader.c

#include "compiler/fzl_token.h"
#include <compiler/fzl_reader.h>
#include <stdbool.h>
#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

// Last error from reader
static char last_err[256] = {0};

fzl_reader_t *fzl_mkreader(char *source)
{
  fzl_reader_ll_t *first_node = malloc(sizeof(fzl_reader_ll_t));
  first_node->token           = NULL;
  first_node->next            = NULL;

  fzl_reader_t *reader = malloc(sizeof(fzl_reader_t));
  reader->first        = first_node;
  reader->last         = first_node;
  reader->source       = source;
  reader->size         = strlen(source);
  reader->position     = 0;
  reader->line_n       = 1;
  reader->line_start   = 0;

  return reader;
}

// --- Read zone ---

// Move cursor forward
static size_t move(fzl_reader_t *reader, size_t len)
{
  if (!reader)
    return 0;

  char *src = reader->source;
  for (size_t i = 0; i < len; i++)
  {
    if (src[reader->position + i] == '\n')
    {
      reader->line_n++;
      reader->line_start = reader->position + i + 1;
    }
  }

  return reader->position += len;
}

// Try parse name
fzl_token_t *parse_name(fzl_reader_t *reader)
{
  size_t lpos = reader->position;
  size_t size = 0;

  do
  {
    lpos++;
    size++;
  } while ((reader->source[lpos] >= 'A' && reader->source[lpos] <= 'Z') ||
           (reader->source[lpos] >= 'a' && reader->source[lpos] <= 'z') ||
           (reader->source[lpos] >= '0' && reader->source[lpos] <= '9') ||
           (reader->source[lpos] == '_'));

  char *value = malloc(size + 1);
  memcpy(value, reader->source + reader->position, size);
  value[size] = '\0';

  size_t start = reader->position;
  move(reader, size);

  fzl_token_t *tok;
  size_t line = reader->line_n;
  size_t pos  = start - reader->line_start;

  if (strcmp(value, "if") == 0)
    tok = fzl_mktoken(FZL_TOKTYPE_IF, line, pos, value, size);
  else if (strcmp(value, "else") == 0)
    tok = fzl_mktoken(FZL_TOKTYPE_ELSE, line, pos, value, size);
  else if (strcmp(value, "true") == 0 || strcmp(value, "false") == 0)
    tok = fzl_mktoken(FZL_TOKTYPE_BOOL_LITERAL, line, pos, value, size);
  else if (strcmp(value, "str") == 0 || strcmp(value, "char") == 0 ||
           strcmp(value, "bool") == 0 || strcmp(value, "uint") == 0 ||
           strcmp(value, "int") == 0 || strcmp(value, "i8") == 0 ||
           strcmp(value, "i16") == 0 || strcmp(value, "i32") == 0 ||
           strcmp(value, "i64") == 0 || strcmp(value, "u8") == 0 ||
           strcmp(value, "u16") == 0 || strcmp(value, "u32") == 0 ||
           strcmp(value, "u64") == 0)
    tok = fzl_mktoken(FZL_TOKTYPE_TYPE, line, pos, value, size);
  else
    tok = fzl_mktoken(FZL_TOKTYPE_NAME, line, pos, value, size);
  free(value);

  return tok;
}

// Try parse string literal
fzl_token_t *parse_string_literal(fzl_reader_t *reader)
{
  size_t lpos = reader->position;
  size_t size = 0;

  if (reader->source[lpos] != '\'' && reader->source[lpos] != '"')
    return NULL;

  char open_symbol  = reader->source[lpos];
  char close_symbol = '\0';

  do
  {
    lpos++;
    size++;

    if (reader->source[lpos] == open_symbol)
      close_symbol = open_symbol;

  } while (reader->source[lpos] != open_symbol && lpos < reader->size);
  size++;

  if (close_symbol != open_symbol)
  {
    snprintf(last_err, sizeof(last_err), "unterminated string at %ld:%ld",
             reader->line_n, reader->position - reader->line_start);
    return NULL;
  }

  if (size == 2)
  {
    fzl_token_t *tok =
        fzl_mktoken(FZL_TOKTYPE_STR_LITERAL, reader->line_n,
                    reader->position - reader->line_start, "", 0);
    move(reader, 2);
    return tok;
  }

  char *value = malloc(size - 2);
  memcpy(value, reader->source + reader->position + 1, size - 2);
  fzl_token_t *tok = fzl_mktoken(FZL_TOKTYPE_STR_LITERAL, reader->line_n,
                                 lpos - reader->line_start, value, size - 2);
  free(value);
  move(reader, size);

  return tok;
}

// Parse numeric literal
fzl_token_t *parse_number(fzl_reader_t *reader)
{
  if (!reader)
    return NULL;

  size_t size = 0;
  size_t lpos = reader->position;

  do
  {
    size++;
    lpos++;
  } while ((reader->source[lpos] >= '0' && reader->source[lpos] <= '9') &&
           lpos < reader->size);

  char *value = malloc(size * sizeof(char) + 1);
  memcpy(value, reader->source + reader->position, size);
  value[size] = '\0';

  fzl_token_t *tok = fzl_mktoken(FZL_TOKTYPE_NUM_LITERAL, reader->line_n,
                                 lpos - reader->line_start, value, size);
  free(value);

  if (!tok)
    return NULL;

  move(reader, size);
  return tok;
}

// Add single-char token
fzl_token_t *add_single(fzl_reader_t *reader, fzl_toktype_t type)
{
  size_t p = move(reader, 1);
  return fzl_mktoken(type, reader->line_n, p - reader->line_start, NULL, 0);
}

// Add double-char token
fzl_token_t *add_double(fzl_reader_t *reader, fzl_toktype_t type)
{
  size_t p = move(reader, 2);
  return fzl_mktoken(type, reader->line_n, p - reader->line_start, NULL, 0);
}

int fzl_read(fzl_reader_t *reader)
{
  if (reader->position >= reader->size)
    return 0;

  do
  {
    size_t p = reader->position;
    fzl_token_t *token;
    char curr = reader->source[p];

    if (curr == '\n' || curr == ' ' || curr == '\r')
    {
      move(reader, 1);
      continue;
    }

    if (curr == '\'' || curr == '"')
      token = parse_string_literal(reader);
    else if (curr == '=' && reader->source[p + 1] == '=')
      token = add_double(reader, FZL_TOKTYPE_EQEQ);
    else if (curr == '>' && reader->source[p + 1] == '=')
      token = add_double(reader, FZL_TOKTYPE_REQ);
    else if (curr == '<' && reader->source[p + 1] == '=')
      token = add_double(reader, FZL_TOKTYPE_LEQ);
    else if (curr == '!' && reader->source[p + 1] == '=')
      token = add_double(reader, FZL_TOKTYPE_NEQ);
    else if (curr == '|' && reader->source[p + 1] == '|')
      token = add_double(reader, FZL_TOKTYPE_DVBAR);
    else if (curr == '&' && reader->source[p + 1] == '&')
      token = add_double(reader, FZL_TOKTYPE_DAMPERSAND);
    else if (curr == '>')
      token = add_single(reader, FZL_TOKTYPE_RANGLE);
    else if (curr == '<')
      token = add_single(reader, FZL_TOKTYPE_LANGLE);
    else if (curr == '!')
      token = add_single(reader, FZL_TOKTYPE_NOT);
    else if (curr == ';')
      token = add_single(reader, FZL_TOKTYPE_SEMICOLON);
    else if (curr == '{')
      token = add_single(reader, FZL_TOKTYPE_LBRACE);
    else if (curr == '}')
      token = add_single(reader, FZL_TOKTYPE_RBRACE);
    else if (curr == '(')
      token = add_single(reader, FZL_TOKTYPE_LPAREN);
    else if (curr == ')')
      token = add_single(reader, FZL_TOKTYPE_RPAREN);
    else if (curr == ',')
      token = add_single(reader, FZL_TOKTYPE_COMMA);
    else if (curr == '*')
      token = add_single(reader, FZL_TOKTYPE_STAR);
    else if (curr == '/')
      token = add_single(reader, FZL_TOKTYPE_SLASH);
    else if (curr == '+')
      token = add_single(reader, FZL_TOKTYPE_PLUS);
    else if (curr == '-')
      token = add_single(reader, FZL_TOKTYPE_MINUS);
    else if (curr == '^')
      token = add_single(reader, FZL_TOKTYPE_CARET);
    else if (curr == '%')
      token = add_single(reader, FZL_TOKTYPE_PERCENT);
    else if (curr == '=')
      token = add_single(reader, FZL_TOKTYPE_EQ);
    else if (curr >= 'A' && curr <= 'Z' || curr >= 'a' && curr <= 'z' ||
             curr == '_')
      token = parse_name(reader);
    else if (curr >= '0' && curr <= '9')
      token = parse_number(reader);
    else
    {
      sprintf(last_err, "undefined token '%c' at %ld:%ld", curr, reader->line_n,
              p - reader->line_start);
      return -1;
    }

    if (!token)
    {
      if (!last_err[0])
        sprintf(last_err, "cannot parse token at %ld:%ld", reader->line_n,
                p - reader->line_start);
      return -1;
    }

    fzl_reader_ll_t *tok_node = malloc(sizeof(fzl_reader_ll_t));
    tok_node->token           = token;

    reader->last->next = tok_node;
    reader->last       = tok_node;
  } while (reader->position < reader->size);

  return 0;
}

// --- Read zone end ---

char *fzl_reader_geterror() { return last_err; }

void free_node_and_next(fzl_reader_ll_t *node)
{
  if (!node)
    return;
  if (node->next)
    free_node_and_next(node->next);

  if (node->token)
    fzl_freetoken(node->token);

  free(node);
}

void fzl_freereader(fzl_reader_t *reader)
{
  if (!reader)
    return;
  if (reader->first)
    free_node_and_next(reader->first);

  free(reader);
}