42sh/src/lexer/lexer.c

#include "lexer.h"

#include <ctype.h>
#include <stdbool.h>
#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include "io_backend/io_backend.h"
#include "utils/string_utils/string_utils.h"

static char *end_last_token;
static ssize_t remaining_chars;
static struct token *last_token;
static struct token *current_token;

/* @brief: sets the current_token to [tok].
 *        this function is called by token_peek().
 */
static void update_current_token(struct token *tok)
{
    current_token = tok;
}

/* @brief: frees the last token and sets it to [tok].
 *        Also sets current_token to NULL.
 *        this function is called by token_pop().
 */
static void update_last_token(struct token *tok)
{
    free_token(&last_token);
    last_token = tok;
}

/* @brief: updates the current position in the stream.
 *          [stream] += [i]
 *        Also frees the last sent token, and sets it to current_token.
 *        Current token is then set to NULL.
 *        This function is called by token_pop().
 */
static void save_state(char *stream, ssize_t i)
{
    remaining_chars -= i;
    end_last_token = stream + i;

    update_last_token(current_token);
    update_current_token(NULL);
}

/* @return: true if a special character from the grammar was found,
 *          false otherwise.
 */
static bool is_special_char(char c)
{
    if (c == EOF)
        return true;

    char special_chars[] = "\n'\"`;#|&\\$(){}<>*";
    return strchr(special_chars, c) != NULL;
}

/* @brief: if a special character is found at [begin],
 * [tok->token_type] is set accordingly
 */
static void set_token_spechar(struct token *tok, char *begin, ssize_t size)
{
    if (size != 1)
        return;
    switch (begin[0])
    {
    case EOF:
        tok->type = TOKEN_EOF;
        break;
    case ';':
        tok->type = TOKEN_SEMICOLON;
        break;
    case '\n':
        tok->type = TOKEN_NEWLINE;
        break;
    case '\'':
        tok->type = TOKEN_QUOTE;
        break;
    case '"':
        tok->type = TOKEN_DOUBLE_QUOTE;
        break;
    case '`':
        tok->type = TOKEN_GRAVE;
        break;
    case '#':
        tok->type = TOKEN_COMMENT;
        break;
    case '|':
        tok->type = TOKEN_PIPE;
        break;
    case '&':
        tok->type = TOKEN_AMPERSAND;
        break;
    case '\\':
        tok->type = TOKEN_BACKSLASH;
        break;
    case '$':
        tok->type = TOKEN_DOLLAR;
        break;
    case '(':
        tok->type = TOKEN_LEFT_PAREN;
        break;
    case ')':
        tok->type = TOKEN_RIGHT_PAREN;
        break;
    case '{':
        tok->type = TOKEN_LEFT_BRACKET;
        break;
    case '}':
        tok->type = TOKEN_RIGHT_BRACKET;
        break;
    case '<':
        tok->type = TOKEN_LESS;
        break;
    case '>':
        tok->type = TOKEN_GREATER;
        break;
    case '*':
        tok->type = TOKEN_STAR;
        break;
    default:
        break;
    }
}

/* @brief: if a keyword is found at [begin],
 * [tok->token_type] is set accordingly
 */
static void set_token_keyword(struct token *tok, char *begin, ssize_t size)
{
    if (tok->type != TOKEN_NULL || size == 0)
        return;
    if (strncmp(begin, "if", size) == 0)
    {
        tok->type = TOKEN_IF;
    }
    else if (strncmp(begin, "fi", size) == 0)
    {
        tok->type = TOKEN_FI;
    }
    else if (strncmp(begin, "then", size) == 0)
    {
        tok->type = TOKEN_THEN;
    }
    else if (strncmp(begin, "else", size) == 0)
    {
        tok->type = TOKEN_ELSE;
    }
    else if (strncmp(begin, "elif", size) == 0)
    {
        tok->type = TOKEN_ELIF;
    }

    // no keywords found.
    if (tok->type == TOKEN_NULL)
        return;

    tok->data = calloc(size + 1, sizeof(char));
    if (tok->data == NULL)
        return;
    strncpy(tok->data, begin, size);
}

/* @brief: if token_type has not yet been set, then it is a TOKEN_WORD
 *         Also allocates the data and fills it.
 */
static void set_token_word(struct token *tok, char *begin, ssize_t size)
{
    if (tok->type == TOKEN_NULL && size != 0)
    {
        tok->type = TOKEN_WORD;
        tok->data = calloc(size + 1, sizeof(char));
        if (tok->data == NULL)
            return;
        strncpy(tok->data, begin, size);
    }
}

struct token *new_token(char *begin, ssize_t size)
{
    struct token *tok = calloc(1, sizeof(struct token));
    if (tok == NULL)
        return NULL;

    set_token_spechar(tok, begin, size);
    set_token_keyword(tok, begin, size);
    set_token_word(tok, begin, size);

    return tok;
}

void free_token(struct token **tok)
{
    if (tok == NULL || *tok == NULL)
        return;
    if ((*tok)->data != NULL)
        free((*tok)->data);
    free(*tok);
    *tok = NULL;
}

char *stream_init(void)
{
    char *stream;

    if (last_token == NULL) // at the begining
    {
        remaining_chars = stream_read(&stream);
    }
    else
    {
        stream = end_last_token;
    }

    char *trimed_stream = trim_blank_left(stream);
    remaining_chars -= trimed_stream - stream;
    stream = trimed_stream;

    return stream;
}

/*
 * @brief: Updates the lexing_mode to LEXER_NORMAL
 *         if the SECOND quote is found at stream[i].
 *         Updates the lexing_mode to the corresponding quote type
 *         if the FIRST quote of any type is found.
 *
 * @return: true if an update was done. false otherwise.
 */
static bool update_lexing_mode(char *stream, ssize_t i,
                               enum lexing_mode *lexing_mode)
{
    enum lexing_mode mode_before_update = *lexing_mode;
    // SECOND quote
    if (*lexing_mode == LEXER_QUOTE && stream[i] == '\'')
        *lexing_mode = LEXER_NORMAL;
    if (*lexing_mode == LEXER_DOUBLE_QUOTE && stream[i] == '"')
        *lexing_mode = LEXER_NORMAL;

    // FIRST quote
    if (*lexing_mode == LEXER_NORMAL)
    {
        if (stream[i] == '"')
            *lexing_mode = LEXER_DOUBLE_QUOTE;

        if (stream[i] == '\'')
            *lexing_mode = LEXER_QUOTE;
    }
    return *lexing_mode != mode_before_update;
}

struct token *peek_token(void)
{
    // we already created the upcoming token during the previous call to peek()
    if (current_token != NULL)
    {
        return current_token;
    }

    char *stream = stream_init();
    ssize_t i = 0;

    // Usefull to know if we are inside a quote or double quote
    enum lexing_mode lexing_mode = LEXER_NORMAL;

    while (i < remaining_chars)
    {
        // true if we didn't encounter a quotes of any type at stream[i]
        if (!update_lexing_mode(stream, i, &lexing_mode))
        {
            if (is_special_char(stream[i]))
            {
                if (i == 0) // where we create spe_char token
                    i++;
                break;
            }
            if (isblank(stream[i]))
            {
                break;
            }
        }
        i++;
    }

    struct token *tok = new_token(stream, i);
    update_current_token(tok);
    return tok;
}

struct token *pop_token(void)
{
    if (current_token != NULL && current_token->type == TOKEN_EOF)
    {
        // we reached end of input, frees all the token still allocated.
        free_token(&last_token);
        free_token(&current_token);
        return NULL;
    }
    char *stream = stream_init();
    ssize_t i = 0;

    // Usefull to know if we are inside a quote or double quote
    enum lexing_mode lexing_mode = LEXER_NORMAL;

    while (i < remaining_chars)
    {
        // true if we didn't encounter a quotes of any type at stream[i]
        if (!update_lexing_mode(stream, i, &lexing_mode))
        {
            if (is_special_char(stream[i]))
            {
                if (i == 0) // where we create spe_char token
                    i++;
                break;
            }
            if (isblank(stream[i]))
            {
                break;
            }
        }
        i++;
    }

    // just in case peek() was not called before poping.
    // (this should never happen)
    if (current_token == NULL)
    {
        current_token = new_token(stream, i);
    }
    save_state(stream, i);

    return last_token;
}