command: relax requirements for unquoted words

Allow most printable characters in unquoted words.  The tokenizer
patch introduced very strict requirements for command parameters -
those were undocumented, and we're reverting the strictness now.

3 files changed, 75 insertions, 6 deletions

diff --git a/src/command.c b/src/command.c
index 7c2743975..f69bdf632 100644
--- a/src/command.c
+++ b/src/command.c
@@ -1959,7 +1959,7 @@ command_process(struct client *client, unsigned num, char *line)
 
 	while (argc < (int)G_N_ELEMENTS(argv) &&
 	       (argv[argc] =
-		tokenizer_next_word_or_string(&line, &error)) != NULL)
+		tokenizer_next_param(&line, &error)) != NULL)
 		++argc;
 
 	/* some error checks; we have to set current_command because
diff --git a/src/tokenizer.c b/src/tokenizer.c
index 635d507df..c1b64f959 100644
--- a/src/tokenizer.c
+++ b/src/tokenizer.c
@@ -90,6 +90,60 @@ tokenizer_next_word(char **input_p, GError **error_r)
 	return word;
 }
 
+static inline bool
+valid_unquoted_char(char ch)
+{
+	return (unsigned char)ch > 0x20 && ch != '"' && ch != '\'';
+}
+
+char *
+tokenizer_next_unquoted(char **input_p, GError **error_r)
+{
+	char *word, *input;
+
+	assert(input_p != NULL);
+	assert(*input_p != NULL);
+
+	word = input = *input_p;
+
+	if (*input == 0)
+		return NULL;
+
+	/* check the first character */
+
+	if (!valid_unquoted_char(*input)) {
+		g_set_error(error_r, tokenizer_quark(), 0,
+			    "Invalid unquoted character");
+		return NULL;
+	}
+
+	/* now iterate over the other characters until we find a
+	   whitespace or end-of-string */
+
+	while (*++input != 0) {
+		if (g_ascii_isspace(*input)) {
+			/* a whitespace: the word ends here */
+			*input = 0;
+			/* skip all following spaces, too */
+			input = g_strchug(input + 1);
+			break;
+		}
+
+		if (!valid_unquoted_char(*input)) {
+			*input_p = input;
+			g_set_error(error_r, tokenizer_quark(), 0,
+				    "Invalid unquoted character");
+			return NULL;
+		}
+	}
+
+	/* end of string: the string is already null-terminated
+	   here */
+
+	*input_p = input;
+	return word;
+}
+
 char *
 tokenizer_next_string(char **input_p, GError **error_r)
 {
@@ -155,7 +209,7 @@ tokenizer_next_string(char **input_p, GError **error_r)
 }
 
 char *
-tokenizer_next_word_or_string(char **input_p, GError **error_r)
+tokenizer_next_param(char **input_p, GError **error_r)
 {
 	assert(input_p != NULL);
 	assert(*input_p != NULL);
@@ -163,5 +217,5 @@ tokenizer_next_word_or_string(char **input_p, GError **error_r)
 	if (**input_p == '"')
 		return tokenizer_next_string(input_p, error_r);
 	else
-		return tokenizer_next_word(input_p, error_r);
+		return tokenizer_next_unquoted(input_p, error_r);
 }
diff --git a/src/tokenizer.h b/src/tokenizer.h
index e0238f0af..ce4c37ccd 100644
--- a/src/tokenizer.h
+++ b/src/tokenizer.h
@@ -37,6 +37,20 @@ char *
 tokenizer_next_word(char **input_p, GError **error_r);
 
 /**
+ * Reads the next unquoted word from the input string.  This function
+ * modifies the input string.
+ *
+ * @param input_p the input string; this function returns a pointer to
+ * the first non-whitespace character of the following token
+ * @param error_r if this function returns NULL and **input_p!=0, it
+ * optionally provides a GError object in this argument
+ * @return a pointer to the null-terminated word, or NULL on error or
+ * end of line
+ */
+char *
+tokenizer_next_unquoted(char **input_p, GError **error_r);
+
+/**
  * Reads the next quoted string from the input string.  A backslash
  * escapes the following character.  This function modifies the input
  * string.
@@ -52,8 +66,9 @@ char *
 tokenizer_next_string(char **input_p, GError **error_r);
 
 /**
- * Reads the next word or quoted string from the input.  This is a
- * wrapper for tokenizer_next_word() and tokenizer_next_string().
+ * Reads the next unquoted word or quoted string from the input.  This
+ * is a wrapper for tokenizer_next_unquoted() and
+ * tokenizer_next_string().
  *
  * @param input_p the input string; this function returns a pointer to
  * the first non-whitespace character of the following token
@@ -63,6 +78,6 @@ tokenizer_next_string(char **input_p, GError **error_r);
  * or end of line
  */
 char *
-tokenizer_next_word_or_string(char **input_p, GError **error_r);
+tokenizer_next_param(char **input_p, GError **error_r);
 
 #endif