yparser: add support for backslashed characters in item value

Backslash character before quotation character is removed during parsing.

This feature is useful for the GeoIP module.

2 files changed, 104 insertions, 81 deletions

diff --git a/src/libknot/yparser/ypbody.c b/src/libknot/yparser/ypbody.c
index f1b13c587..73b77f52f 100644
--- a/src/libknot/yparser/ypbody.c
+++ b/src/libknot/yparser/ypbody.c
@@ -27,24 +27,25 @@
 
 static const char _yparser_actions[] = {
 	0, 1, 0, 1, 1, 1, 2, 1,
-	3, 1, 4, 1, 5, 1, 7, 1,
-	8, 1, 9, 1, 10, 1, 11, 1,
-	14, 2, 1, 0, 2, 1, 2, 2,
-	2, 0, 2, 3, 4, 2, 5, 0,
-	2, 6, 7, 2, 11, 12, 2, 13,
-	11, 3, 1, 2, 0, 3, 1, 6,
-	7, 3, 1, 11, 12, 3, 1, 13,
-	11, 3, 2, 6, 7, 3, 2, 11,
-	12, 3, 2, 13, 11, 4, 1, 2,
-	6, 7, 4, 1, 2, 11, 12, 4,
-	1, 2, 13, 11
+	3, 1, 4, 1, 6, 1, 8, 1,
+	9, 1, 10, 1, 11, 1, 12, 1,
+	15, 2, 1, 0, 2, 1, 2, 2,
+	2, 0, 2, 3, 4, 2, 5, 4,
+	2, 6, 0, 2, 7, 8, 2, 12,
+	13, 2, 14, 12, 3, 1, 2, 0,
+	3, 1, 7, 8, 3, 1, 12, 13,
+	3, 1, 14, 12, 3, 2, 7, 8,
+	3, 2, 12, 13, 3, 2, 14, 12,
+	4, 1, 2, 7, 8, 4, 1, 2,
+	12, 13, 4, 1, 2, 14, 12
 };
 
 static const unsigned char _yparser_key_offsets[] = {
 	0, 0, 13, 15, 16, 25, 36, 38,
-	39, 49, 60, 71, 73, 76, 88, 91,
-	94, 98, 100, 110, 118, 121, 124, 127,
-	130, 141, 154, 167, 180
+	39, 49, 60, 71, 73, 76, 89, 93,
+	96, 100, 102, 105, 115, 124, 127, 130,
+	133, 137, 140, 143, 146, 157, 170, 183,
+	196
 };
 
 static const char _yparser_trans_keys[] = {
@@ -58,13 +59,15 @@ static const char _yparser_trans_keys[] = {
 	45, 90, 94, 126, 32, 58, 92, 45,
 	46, 48, 57, 65, 90, 97, 122, 32,
 	58, 10, 13, 32, 10, 13, 32, 34,
-	35, 91, 33, 43, 45, 92, 94, 126,
-	34, 32, 126, 10, 13, 32, 10, 13,
-	32, 35, 10, 13, 32, 33, 34, 92,
-	36, 43, 45, 90, 94, 126, 32, 33,
-	44, 93, 36, 90, 92, 126, 32, 44,
-	93, 10, 13, 32, 34, 32, 126, 32,
-	44, 93, 32, 58, 92, 45, 46, 48,
+	35, 91, 92, 33, 43, 45, 90, 94,
+	126, 34, 92, 32, 126, 10, 13, 32,
+	10, 13, 32, 35, 10, 13, 34, 32,
+	126, 32, 33, 34, 92, 36, 43, 45,
+	90, 94, 126, 32, 33, 44, 92, 93,
+	36, 90, 94, 126, 32, 44, 93, 10,
+	13, 32, 34, 32, 126, 34, 92, 32,
+	126, 32, 44, 93, 34, 32, 126, 34,
+	32, 126, 32, 58, 92, 45, 46, 48,
 	57, 65, 90, 97, 122, 10, 13, 32,
 	35, 45, 46, 92, 48, 57, 65, 90,
 	97, 122, 10, 13, 32, 35, 45, 46,
@@ -77,23 +80,26 @@ static const char _yparser_trans_keys[] = {
 
 static const char _yparser_single_lengths[] = {
 	0, 7, 2, 1, 3, 3, 2, 1,
-	4, 5, 3, 2, 3, 6, 1, 3,
-	4, 2, 4, 4, 3, 3, 1, 3,
-	3, 7, 7, 7, 7
+	4, 5, 3, 2, 3, 7, 2, 3,
+	4, 2, 1, 4, 5, 3, 3, 1,
+	2, 3, 1, 1, 3, 7, 7, 7,
+	7
 };
 
 static const char _yparser_range_lengths[] = {
 	0, 3, 0, 0, 3, 4, 0, 0,
 	3, 3, 4, 0, 0, 3, 1, 0,
-	0, 0, 3, 2, 0, 0, 1, 0,
-	4, 3, 3, 3, 3
+	0, 0, 1, 3, 2, 0, 0, 1,
+	1, 0, 1, 1, 4, 3, 3, 3,
+	3
 };
 
 static const unsigned char _yparser_index_offsets[] = {
 	0, 0, 11, 14, 16, 23, 31, 34,
-	36, 44, 53, 61, 64, 68, 78, 81,
-	85, 90, 93, 101, 108, 112, 116, 119,
-	123, 131, 142, 153, 164
+	36, 44, 53, 61, 64, 68, 79, 83,
+	87, 92, 95, 98, 106, 114, 118, 122,
+	125, 129, 133, 136, 139, 147, 158, 169,
+	180
 };
 
 static const char _yparser_indicies[] = {
@@ -101,62 +107,67 @@ static const char _yparser_indicies[] = {
 	6, 6, 0, 1, 2, 4, 7, 0,
 	7, 8, 8, 8, 8, 8, 0, 9,
 	11, 10, 10, 10, 10, 10, 0, 12,
-	13, 0, 14, 0, 14, 15, 16, 15,
-	15, 15, 15, 0, 17, 18, 19, 20,
-	20, 20, 20, 20, 0, 21, 23, 22,
-	22, 22, 22, 22, 0, 24, 25, 0,
-	26, 27, 28, 0, 26, 27, 28, 16,
-	29, 30, 15, 15, 15, 0, 32, 31,
-	0, 17, 18, 19, 0, 26, 27, 33,
-	29, 0, 26, 27, 29, 30, 34, 35,
-	34, 34, 34, 34, 0, 36, 37, 38,
-	39, 37, 37, 0, 40, 30, 41, 0,
-	26, 27, 33, 0, 43, 42, 0, 36,
-	38, 39, 0, 44, 46, 45, 45, 45,
-	45, 45, 0, 1, 2, 3, 4, 5,
-	47, 47, 47, 47, 47, 0, 48, 49,
-	50, 51, 52, 53, 53, 53, 53, 53,
-	0, 54, 55, 56, 57, 58, 59, 59,
-	59, 59, 59, 0, 60, 61, 62, 63,
-	64, 65, 65, 65, 65, 65, 0, 0
+	13, 0, 14, 0, 14, 15, 16, 17,
+	15, 15, 15, 0, 18, 19, 20, 21,
+	22, 21, 21, 21, 0, 23, 25, 24,
+	24, 24, 24, 24, 0, 26, 27, 0,
+	28, 29, 30, 0, 28, 29, 30, 16,
+	31, 32, 17, 15, 15, 15, 0, 34,
+	35, 33, 0, 18, 19, 20, 0, 28,
+	29, 36, 31, 0, 28, 29, 31, 37,
+	33, 0, 32, 38, 39, 40, 38, 38,
+	38, 0, 41, 42, 43, 44, 45, 42,
+	42, 0, 46, 32, 47, 0, 28, 29,
+	36, 0, 48, 42, 0, 50, 51, 49,
+	0, 41, 43, 45, 0, 52, 49, 0,
+	53, 21, 0, 54, 56, 55, 55, 55,
+	55, 55, 0, 1, 2, 3, 4, 5,
+	57, 57, 57, 57, 57, 0, 58, 59,
+	60, 61, 62, 63, 63, 63, 63, 63,
+	0, 64, 65, 66, 67, 68, 69, 69,
+	69, 69, 69, 0, 70, 71, 72, 73,
+	74, 75, 75, 75, 75, 75, 0, 0
 };
 
 static const char _yparser_trans_targs[] = {
-	0, 26, 27, 1, 2, 3, 24, 4,
+	0, 30, 31, 1, 2, 3, 28, 4,
 	5, 6, 5, 7, 6, 7, 8, 9,
-	14, 28, 25, 16, 9, 11, 10, 12,
-	11, 12, 28, 25, 13, 17, 18, 14,
-	15, 16, 19, 22, 20, 19, 18, 21,
-	20, 21, 22, 23, 11, 24, 12, 10,
-	26, 27, 1, 2, 3, 10, 26, 27,
-	1, 2, 3, 10, 26, 27, 1, 2,
-	3, 10
+	14, 27, 32, 29, 16, 9, 27, 11,
+	10, 12, 11, 12, 32, 29, 13, 17,
+	19, 14, 15, 18, 16, 14, 20, 24,
+	23, 21, 20, 19, 23, 22, 21, 22,
+	20, 24, 25, 26, 24, 9, 11, 28,
+	12, 10, 30, 31, 1, 2, 3, 10,
+	30, 31, 1, 2, 3, 10, 30, 31,
+	1, 2, 3, 10
 };
 
 static const char _yparser_trans_actions[] = {
-	23, 1, 0, 43, 0, 46, 40, 21,
-	40, 19, 13, 19, 0, 0, 0, 34,
-	7, 37, 11, 11, 9, 15, 13, 15,
-	0, 0, 1, 0, 0, 0, 0, 9,
-	0, 0, 34, 7, 11, 9, 11, 11,
-	0, 0, 9, 0, 17, 13, 17, 40,
-	49, 28, 82, 28, 87, 77, 31, 5,
-	69, 5, 73, 65, 25, 3, 57, 3,
-	61, 53
+	23, 1, 0, 46, 0, 49, 43, 21,
+	43, 19, 13, 19, 0, 0, 0, 34,
+	7, 34, 40, 11, 11, 9, 9, 15,
+	13, 15, 0, 0, 1, 0, 0, 0,
+	0, 9, 0, 9, 0, 37, 34, 7,
+	34, 11, 9, 11, 9, 11, 0, 0,
+	37, 9, 0, 9, 37, 37, 17, 13,
+	17, 43, 52, 28, 85, 28, 90, 80,
+	31, 5, 72, 5, 76, 68, 25, 3,
+	60, 3, 64, 56
 };
 
 static const char _yparser_eof_actions[] = {
 	0, 23, 23, 23, 23, 23, 23, 23,
 	23, 23, 23, 23, 23, 23, 23, 23,
 	23, 23, 23, 23, 23, 23, 23, 23,
-	23, 0, 28, 5, 3
+	23, 23, 23, 23, 23, 0, 28, 5,
+	3
 };
 
 
 
 
 
-int _yp_start_state = 25;
+int _yp_start_state = 29;
 
 int _yp_parse(
 	yp_parser_t *parser)
@@ -294,6 +305,11 @@ _match:
 	break;
 	case 5:
 	{
+		parser->data_len--;
+	}
+	break;
+	case 6:
+	{
 		// Return if a value parsed.
 		parser->data[parser->data_len] = '\0';
 		parser->processed = true;
@@ -301,7 +317,7 @@ _match:
 		{p++; goto _out; }
 	}
 	break;
-	case 6:
+	case 7:
 	{
 		if (indent > 0 && parser->indent > 0 &&
 		    indent != parser->indent) {
@@ -313,7 +329,7 @@ _match:
 		parser->event = YP_ENULL;
 	}
 	break;
-	case 7:
+	case 8:
 	{
 		if (parser->key_len >= sizeof(parser->key) - 1) {
 			return KNOT_ESPACE;
@@ -321,7 +337,7 @@ _match:
 		parser->key[parser->key_len++] = (*p);
 	}
 	break;
-	case 8:
+	case 9:
 	{
 		parser->key[parser->key_len] = '\0';
 		parser->indent = 0;
@@ -329,14 +345,14 @@ _match:
 		parser->event = YP_EKEY0;
 	}
 	break;
-	case 9:
+	case 10:
 	{
 		parser->key[parser->key_len] = '\0';
 		parser->indent = indent;
 		parser->event = YP_EKEY1;
 	}
 	break;
-	case 10:
+	case 11:
 	{
 		parser->key[parser->key_len] = '\0';
 		parser->indent = indent;
@@ -344,17 +360,17 @@ _match:
 		parser->event = YP_EID;
 	}
 	break;
-	case 11:
+	case 12:
 	{
 		indent++;
 	}
 	break;
-	case 12:
+	case 13:
 	{
 		id_pos++;
 	}
 	break;
-	case 13:
+	case 14:
 	{
 		if (id_pos > 0 && parser->id_pos > 0 &&
 		    id_pos != parser->id_pos) {
@@ -363,7 +379,7 @@ _match:
 		parser->indent = 0;
 	}
 	break;
-	case 14:
+	case 15:
 	{
 		switch ((*p)) {
 		case '\t':
@@ -401,7 +417,7 @@ _again:
 		id_pos = 0;
 	}
 	break;
-	case 14:
+	case 15:
 	{
 		switch ((*p)) {
 		case '\t':
diff --git a/src/libknot/yparser/ypbody.rl b/src/libknot/yparser/ypbody.rl
index 9e1f9830f..b8f1a08fd 100644
--- a/src/libknot/yparser/ypbody.rl
+++ b/src/libknot/yparser/ypbody.rl
@@ -66,6 +66,9 @@
 		}
 		parser->data[parser->data_len++] = fc;
 	}
+	action _item_unbackslash {
+		parser->data_len--;
+	}
 	action _item_data_exit {
 		// Return if a value parsed.
 		parser->data[parser->data_len] = '\0';
@@ -73,16 +76,20 @@
 		found = true;
 		fbreak;
 	}
+	backslash_char = '\\';
 	quote_char = '\"';
 	list_char = [\[,\]];
 	data_char =
-		(ascii - space - cntrl - quote_char - sep_char -
-		 comment_char - list_char
+		( (ascii - space - cntrl - quote_char - sep_char -
+		   comment_char - list_char - backslash_char)
+		| (backslash_char . (quote_char) >_item_unbackslash)
+		| (backslash_char . (32..126 - quote_char))
 		) $_item_data;
 	data_str_char =
-		(data_char | sep_char | comment_char | list_char
-		) $_item_data;
-	data_str = (quote_char . data_str_char* . quote_char);
+		( (data_char)
+		| (sep_char | comment_char | list_char) $_item_data
+		);
+	data_str = (quote_char . data_str_char* <: quote_char);
 	item_data = (data_char+ | data_str) >_item_data_init %_item_data_exit;
 	item_data_plus = item_data . ((sep? . ',' . sep?) . item_data)*;
 	item_data_list = '\[' . sep? . item_data_plus . sep? . '\]';