wordsplit: handle single quotes

src/parser/wordsplit/wordsplit.c

@@ -6,14 +6,14 @@
 /*   By: khais <marvin@42.fr>                       +#+  +:+       +#+        */
 /*                                                +#+#+#+#+#+   +#+           */
 /*   Created: 2025/02/13 17:02:32 by khais             #+#    #+#             */
-/*   Updated: 2025/02/14 18:06:44 by khais            ###   ########.fr       */
+/*   Updated: 2025/02/14 18:45:45 by khais            ###   ########.fr       */
 /*                                                                            */
 /* ************************************************************************** */
 
 #include "wordsplit.h"
-#include "libft.h"
 #include "../../buffer/buffer.h"
 #include "../matchers/blank.h"
+#include <stdlib.h>
 
 static t_wordlist	*delimit(t_wordlist *wordlist, t_buffer **token, bool *currently_in_word)
 {
@@ -28,20 +28,25 @@ static t_wordlist	*delimit(t_wordlist *wordlist, t_buffer **token, bool *current
 
 static t_buffer	*push_char(t_buffer *token, char c)
 {
+	if (token == NULL)
+		token = ft_buffer_new();
 	return (ft_buffer_pushchar(token, c));
 }
 
 
-static t_buffer	*new_word(char c, bool *currently_in_word)
+static t_buffer	*new_word(t_buffer *token, char c, bool *currently_in_word)
 {
-	t_buffer	*token;
+	token = push_char(token, c);
-
-	token = ft_buffer_new();
-	ft_buffer_pushchar(token, c);
 	(*currently_in_word) = true;
 	return (token);
 }
 
+static bool	quote_flip(t_buffer **token, char c, bool currently_in_quotes)
+{
+	(*token) = push_char((*token), c);
+	return (!currently_in_quotes);
+}
+
 /*
 ** split a string into words, respecting quotes etc.
 **
@@ -58,21 +63,25 @@ t_wordlist	*minishell_wordsplit(char *original)
 	t_wordlist	*wordlist;
 	t_buffer	*token;
 	bool		currently_in_word;
+	bool		currently_in_quotes;
 
 	idx = 0;
 	wordlist = NULL;
 	token = NULL;
 	currently_in_word = false;
+	currently_in_quotes = false;
 	while (true)
 	{
 		// If the end of input is recognized, the current token (if any) shall
 		// be delimited.
 		if (original[idx] == '\0')
 			wordlist = delimit(wordlist, &token, &currently_in_word);
+		else if (original[idx] == '\'')
+			currently_in_quotes = quote_flip(&token, original[idx], currently_in_quotes);
 		// If the current character is an unquoted <blank>, any token containing
 		// the previous character is delimited and the current character shall
 		// be discarded.
-		else if (is_blank(original[idx]))
+		else if (is_blank(original[idx]) && !currently_in_quotes)
 			wordlist = delimit(wordlist, &token, &currently_in_word);
 		// If the previous character was part of a word, the current character
 		// shall be appended to that word.
@@ -80,7 +89,7 @@ t_wordlist	*minishell_wordsplit(char *original)
 			token = push_char(token, original[idx]);
 		// The current character is used as the start of a new word.
 		else
-			token = new_word(original[idx], &currently_in_word);
+			token = new_word(token, original[idx], &currently_in_word);
 		if (original[idx] == '\0')
 			break ;
 		idx++;

tests/word_splitting.c

@@ -6,7 +6,7 @@
 /*   By: khais <marvin@42.fr>                       +#+  +:+       +#+        */
 /*                                                +#+#+#+#+#+   +#+           */
 /*   Created: 2025/02/13 15:17:56 by khais             #+#    #+#             */
-/*   Updated: 2025/02/14 16:16:14 by khais            ###   ########.fr       */
+/*   Updated: 2025/02/14 17:37:46 by khais            ###   ########.fr       */
 /*                                                                            */
 /* ************************************************************************** */
 
@@ -54,9 +54,21 @@ void	test_wordsplit_multiword(void)
 	wordlist_destroy(words);
 }
 
+void	test_wordsplit_multiword_with_single_quotes(void)
+{
+	t_wordlist	*words;
+
+	words = minishell_wordsplit("\t  echo\t' \t The  file is named    $MYFILE  ' \t");
+	assert_strequal("echo", wordlist_get(words, 0)->word);
+	assert_strequal("' \t The  file is named    $MYFILE  '", wordlist_get(words, 1)->word);
+	assert(NULL == wordlist_get(words, 2));
+	wordlist_destroy(words);
+}
+
 int main(void) {
 	test_wordsplit_singleword();
 	test_wordsplit_singleword_with_blanks();
 	test_wordsplit_multiword();
+	test_wordsplit_multiword_with_single_quotes();
 	return (0);
 }