yparser: remove relaxed formatting rules

src/libknot/internal/yparser/yparser.h

@@ -46,6 +46,10 @@ typedef struct {
 	int cs;
 	/*! Indication if the current item was already processed. */
 	bool processed;
+	/*! Current block indentation. */
+	size_t indent;
+	/*! Last id dash position. */
+	size_t id_pos;
 
 	/*! Input parameters. */
 	struct {

src/libknot/internal/yparser/ypbody.c

@@ -28,126 +28,110 @@
 static const char _yparser_actions[] = {
 	0, 1, 0, 1, 1, 1, 2, 1, 
 	3, 1, 4, 1, 6, 1, 7, 1, 
-	8, 1, 9, 1, 10, 2, 1, 0, 
-	2, 2, 3, 2, 4, 0, 2, 5, 
-	6, 3, 1, 5, 6
+	8, 1, 9, 1, 10, 1, 13, 2, 
+	1, 0, 2, 2, 3, 2, 4, 0, 
+	2, 5, 6, 2, 10, 11, 2, 12, 
+	10, 3, 1, 5, 6, 3, 1, 10, 
+	11, 3, 1, 12, 10
 };
 
 static const unsigned char _yparser_key_offsets[] = {
-	0, 0, 12, 13, 22, 33, 37, 49, 
-	62, 67, 72, 77, 88, 92, 105, 117, 
-	129, 134, 140, 152, 164, 174, 185, 190, 
-	195, 206, 218
+	0, 0, 10, 11, 12, 19, 28, 30, 
+	31, 41, 51, 54, 57, 59, 68, 70, 
+	71, 81, 91, 99, 102, 104, 107, 110, 
+	119, 121, 123, 134, 144
 };
 
 static const char _yparser_trans_keys[] = {
-	9, 10, 13, 32, 35, 45, 48, 57, 
-	65, 90, 97, 122, 10, 9, 13, 32, 
-	48, 57, 65, 90, 97, 122, 9, 13, 
-	32, 45, 58, 48, 57, 65, 90, 97, 
-	122, 9, 13, 32, 58, 9, 13, 32, 
+	10, 32, 35, 45, 48, 57, 65, 90, 
+	97, 122, 10, 32, 32, 48, 57, 65, 
+	90, 97, 122, 32, 45, 58, 48, 57, 
+	65, 90, 97, 122, 32, 58, 32, 32, 
 	33, 34, 92, 36, 43, 45, 90, 94, 
-	126, 9, 10, 13, 32, 33, 35, 92, 
-	36, 43, 45, 90, 94, 126, 9, 10, 
-	13, 32, 35, 9, 13, 34, 32, 126, 
-	9, 10, 13, 32, 35, 9, 13, 32, 
-	45, 58, 48, 57, 65, 90, 97, 122, 
-	9, 13, 32, 58, 9, 10, 13, 32, 
-	34, 35, 91, 33, 43, 45, 92, 94, 
-	126, 9, 10, 13, 32, 33, 35, 44, 
-	92, 36, 90, 94, 126, 9, 10, 13, 
-	32, 34, 35, 44, 92, 33, 90, 94, 
-	126, 9, 13, 34, 32, 126, 9, 10, 
-	13, 32, 35, 44, 9, 13, 32, 33, 
-	34, 92, 36, 43, 45, 90, 94, 126, 
-	9, 13, 32, 33, 34, 92, 36, 43, 
-	45, 90, 94, 126, 9, 13, 32, 33, 
-	44, 93, 36, 90, 92, 126, 9, 13, 
-	32, 33, 34, 44, 93, 36, 90, 92, 
-	126, 9, 13, 34, 32, 126, 9, 13, 
-	32, 44, 93, 9, 13, 32, 45, 58, 
-	48, 57, 65, 90, 97, 122, 9, 10, 
-	13, 32, 35, 45, 48, 57, 65, 90, 
-	97, 122, 9, 10, 13, 32, 35, 45, 
-	48, 57, 65, 90, 97, 122, 0
+	126, 10, 32, 33, 92, 36, 43, 45, 
+	90, 94, 126, 10, 32, 35, 34, 32, 
+	126, 10, 32, 32, 45, 58, 48, 57, 
+	65, 90, 97, 122, 32, 58, 32, 32, 
+	33, 34, 91, 36, 43, 45, 92, 94, 
+	126, 32, 33, 34, 92, 36, 43, 45, 
+	90, 94, 126, 32, 33, 44, 93, 36, 
+	90, 92, 126, 32, 44, 93, 10, 32, 
+	34, 32, 126, 32, 44, 93, 32, 45, 
+	58, 48, 57, 65, 90, 97, 122, 32, 
+	58, 10, 32, 10, 32, 34, 35, 91, 
+	33, 43, 45, 92, 94, 126, 10, 32, 
+	35, 45, 48, 57, 65, 90, 97, 122, 
+	10, 32, 35, 45, 48, 57, 65, 90, 
+	97, 122, 0
 };
 
 static const char _yparser_single_lengths[] = {
-	0, 6, 1, 3, 5, 4, 6, 7, 
-	5, 3, 5, 5, 4, 7, 8, 8, 
-	3, 6, 6, 6, 6, 7, 3, 5, 
-	5, 6, 6
+	0, 4, 1, 1, 1, 3, 2, 1, 
+	4, 4, 3, 1, 2, 3, 2, 1, 
+	4, 4, 4, 3, 2, 1, 3, 3, 
+	2, 2, 5, 4, 4
 };
 
 static const char _yparser_range_lengths[] = {
-	0, 3, 0, 3, 3, 0, 3, 3, 
-	0, 1, 0, 3, 0, 3, 2, 2, 
-	1, 0, 3, 3, 2, 2, 1, 0, 
-	3, 3, 3
+	0, 3, 0, 0, 3, 3, 0, 0, 
+	3, 3, 0, 1, 0, 3, 0, 0, 
+	3, 3, 2, 0, 0, 1, 0, 3, 
+	0, 0, 3, 3, 3
 };
 
 static const unsigned char _yparser_index_offsets[] = {
-	0, 0, 10, 12, 19, 28, 33, 43, 
-	54, 60, 65, 71, 80, 85, 96, 107, 
-	118, 123, 130, 140, 150, 159, 169, 174, 
-	180, 189, 199
+	0, 0, 8, 10, 12, 17, 24, 27, 
+	29, 37, 45, 49, 52, 55, 62, 65, 
+	67, 75, 83, 90, 94, 97, 100, 104, 
+	111, 114, 117, 126, 134
 };
 
 static const char _yparser_indicies[] = {
-	1, 2, 1, 1, 3, 4, 5, 5, 
-	5, 0, 2, 3, 4, 4, 4, 6, 
-	6, 6, 0, 7, 7, 7, 8, 9, 
-	8, 8, 8, 0, 10, 10, 10, 11, 
-	0, 11, 11, 11, 12, 13, 12, 12, 
-	12, 12, 0, 14, 15, 14, 14, 16, 
-	17, 16, 16, 16, 16, 0, 18, 2, 
-	18, 18, 3, 0, 19, 19, 20, 19, 
-	0, 14, 15, 14, 14, 17, 0, 21, 
-	21, 21, 22, 23, 22, 22, 22, 0, 
-	24, 24, 24, 25, 0, 25, 2, 25, 
-	25, 27, 3, 28, 26, 26, 26, 0, 
-	29, 15, 29, 29, 30, 17, 31, 30, 
-	30, 30, 0, 32, 2, 32, 32, 27, 
-	3, 33, 26, 26, 26, 0, 34, 34, 
-	35, 34, 0, 29, 15, 29, 29, 17, 
-	31, 0, 33, 33, 33, 26, 27, 26, 
-	26, 26, 26, 0, 28, 28, 28, 36, 
-	37, 36, 36, 36, 36, 0, 38, 38, 
-	38, 39, 40, 14, 39, 39, 0, 41, 
-	41, 41, 36, 37, 28, 18, 36, 36, 
-	0, 42, 42, 43, 42, 0, 38, 38, 
-	38, 40, 14, 0, 44, 44, 44, 45, 
-	46, 45, 45, 45, 0, 1, 2, 1, 
-	1, 3, 4, 47, 47, 47, 0, 48, 
-	49, 48, 48, 50, 51, 52, 52, 52, 
-	0, 0
+	1, 2, 3, 4, 5, 5, 5, 0, 
+	1, 3, 6, 0, 6, 7, 7, 7, 
+	0, 8, 9, 10, 9, 9, 9, 0, 
+	11, 12, 0, 13, 0, 13, 14, 15, 
+	14, 14, 14, 14, 0, 16, 17, 18, 
+	18, 18, 18, 18, 0, 1, 19, 3, 
+	0, 21, 20, 0, 16, 17, 0, 22, 
+	23, 24, 23, 23, 23, 0, 25, 26, 
+	0, 27, 0, 27, 14, 15, 28, 14, 
+	14, 14, 0, 28, 29, 30, 29, 29, 
+	29, 29, 0, 31, 32, 33, 34, 32, 
+	32, 0, 35, 28, 36, 0, 1, 19, 
+	0, 38, 37, 0, 31, 33, 34, 0, 
+	39, 40, 41, 40, 40, 40, 0, 42, 
+	43, 0, 1, 44, 0, 1, 44, 15, 
+	3, 28, 14, 14, 14, 0, 1, 2, 
+	3, 4, 45, 45, 45, 0, 46, 47, 
+	48, 49, 50, 50, 50, 0, 0
 };
 
 static const char _yparser_trans_targs[] = {
-	0, 1, 26, 2, 3, 24, 4, 5, 
-	4, 6, 5, 6, 7, 9, 8, 26, 
-	7, 2, 8, 9, 10, 12, 11, 13, 
-	12, 13, 14, 16, 19, 15, 14, 18, 
-	15, 18, 16, 17, 20, 22, 21, 20, 
-	19, 21, 22, 23, 12, 24, 13, 11, 
-	1, 26, 2, 3, 11
+	0, 28, 1, 2, 3, 13, 4, 5, 
+	6, 5, 7, 6, 7, 8, 9, 11, 
+	28, 10, 9, 10, 11, 12, 14, 13, 
+	15, 14, 15, 16, 17, 18, 21, 19, 
+	18, 17, 20, 19, 20, 21, 22, 24, 
+	23, 25, 24, 25, 26, 23, 28, 1, 
+	2, 3, 23
 };
 
 static const char _yparser_trans_actions[] = {
-	19, 0, 1, 0, 0, 30, 30, 17, 
-	11, 17, 0, 0, 24, 5, 9, 27, 
-	7, 9, 0, 7, 0, 13, 11, 13, 
-	0, 0, 24, 5, 0, 9, 7, 9, 
-	0, 0, 7, 0, 24, 5, 9, 7, 
-	9, 0, 7, 0, 15, 11, 15, 30, 
-	3, 21, 3, 3, 33
+	21, 1, 35, 0, 38, 32, 19, 32, 
+	17, 11, 17, 0, 0, 0, 26, 5, 
+	29, 9, 7, 0, 7, 0, 15, 11, 
+	15, 0, 0, 0, 0, 26, 5, 9, 
+	7, 9, 9, 0, 0, 7, 0, 13, 
+	11, 13, 0, 0, 0, 32, 23, 45, 
+	3, 49, 41
 };
 
 static const char _yparser_eof_actions[] = {
-	0, 19, 19, 19, 19, 19, 19, 19, 
-	19, 19, 19, 19, 19, 19, 19, 19, 
-	19, 19, 19, 19, 19, 19, 19, 19, 
-	19, 0, 3
+	0, 21, 21, 21, 21, 21, 21, 21, 
+	21, 21, 21, 21, 21, 21, 21, 21, 
+	21, 21, 21, 21, 21, 21, 21, 21, 
+	21, 21, 21, 0, 3
 };
 
 
@@ -155,7 +139,7 @@ static const char _yparser_eof_actions[] = {
 
 
 int _yp_start_state = 
-25
+27
 ;
 
 int _yp_parse(
@@ -164,6 +148,10 @@ int _yp_parse(
 	// Parser input limits (Ragel internals).
 	const char *p, *pe, *eof;
 
+	// Current item indent.
+	size_t indent = 0;
+	// Current id dash position.
+	size_t id_pos = 0;
 	// Indicates if the current parsing step contains an item.
 	bool found = false;
 
@@ -293,6 +281,10 @@ _match:
 	break;
 	case 5:
 	{
+		if (indent > 0 && parser->indent > 0 &&
+		    indent != parser->indent) {
+			return KNOT_EPARSEFAIL;
+		}
 		parser->processed = false;
 		parser->key_len = 0;
 		parser->data_len = 0;
@@ -310,22 +302,45 @@ _match:
 	case 7:
 	{
 		parser->key[parser->key_len] = '\0';
+		parser->indent = 0;
 		parser->event = YP_EKEY0;
 	}
 	break;
 	case 8:
 	{
 		parser->key[parser->key_len] = '\0';
+		parser->indent = indent;
 		parser->event = YP_EKEY1;
 	}
 	break;
 	case 9:
 	{
 		parser->key[parser->key_len] = '\0';
+		parser->indent = indent;
+		parser->id_pos = id_pos;
 		parser->event = YP_EID;
 	}
 	break;
 	case 10:
+	{
+		indent++;
+	}
+	break;
+	case 11:
+	{
+		id_pos++;
+	}
+	break;
+	case 12:
+	{
+		if (id_pos > 0 && parser->id_pos > 0 &&
+		    id_pos != parser->id_pos) {
+			return KNOT_EPARSEFAIL;
+		}
+		parser->indent = 0;
+	}
+	break;
+	case 13:
 	{
 		return KNOT_EPARSEFAIL;
 	}
@@ -352,7 +367,7 @@ _again:
 		parser->processed = false;
 	}
 	break;
-	case 10:
+	case 13:
 	{
 		return KNOT_EPARSEFAIL;
 	}

src/libknot/internal/yparser/ypbody.rl

@@ -45,10 +45,11 @@
 	comment      = comment_char . (^newline_char)*;
 
 	# White space processing.
-	sep_char = [ \t\r];
+	sep_char = ' ';
 	sep = sep_char+;
 
 	blank = (sep? :> comment?). newline_char;
+	rest  = ((sep :> comment) | sep?). newline_char;
 
 	# Data processing.
 	action _item_data_init {
@@ -78,15 +79,15 @@
 		) $_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 . ',' . sep) | (sep . ',') | (',' . sep)) .
-		 item_data)*;
-	item_data_list =
-		('\[' . sep? . item_data_plus . sep? . '\]') | item_data_plus;
+	item_data_plus = item_data . ((sep? . ',' . sep?) . item_data)*;
+	item_data_list = '\[' . sep? . item_data_plus . sep? . '\]';
 
 	# Key processing.
 	action _key_init {
+		if (indent > 0 && parser->indent > 0 &&
+		    indent != parser->indent) {
+			return KNOT_EPARSEFAIL;
+		}
 		parser->processed = false;
 		parser->key_len = 0;
 		parser->data_len = 0;
@@ -100,25 +101,41 @@
 	}
 	action _key0_exit {
 		parser->key[parser->key_len] = '\0';
+		parser->indent = 0;
 		parser->event = YP_EKEY0;
 	}
 	action _key1_exit {
 		parser->key[parser->key_len] = '\0';
+		parser->indent = indent;
 		parser->event = YP_EKEY1;
 	}
 	action _id_exit {
 		parser->key[parser->key_len] = '\0';
+		parser->indent = indent;
+		parser->id_pos = id_pos;
 		parser->event = YP_EID;
 	}
-	key_data_sep = sep? . ":" . sep?;
-	key_name = alnum . ("-" | alnum)*;
-	key0 =        key_name              >_key_init $_key %_key0_exit;
-	key1 = sep .  key_name              >_key_init $_key %_key1_exit;
-	id   = sep? . "-" . sep? . key_name >_key_init $_key %_id_exit;
-	item = (((key0 . key_data_sep . item_data_list?)
-	        |(key1 . key_data_sep . item_data_list?)
-	        |(id   . key_data_sep . item_data )
-	        ) . blank
+	action _indent {
+		indent++;
+	}
+	action _id {
+		id_pos++;
+	}
+	action _dash_init {
+		if (id_pos > 0 && parser->id_pos > 0 &&
+		    id_pos != parser->id_pos) {
+			return KNOT_EPARSEFAIL;
+		}
+		parser->indent = 0;
+	}
+	key_name = (alnum . ("-" | alnum)*) >_key_init $_key;
+	key0 =                                                  key_name %_key0_exit;
+	key1 =   sep $_indent .                                 key_name %_key1_exit;
+	id   = ((sep $_id)? . "-" >_dash_init . sep) $_indent . key_name %_id_exit;
+	item = (((key0 . sep? . ":" . (sep . (item_data_list | item_data))?)
+	        |(key1 . sep? . ":" .  sep . (item_data_list | item_data))
+	        |(id   . sep? . ":" .  sep . item_data)
+	        ) . rest
 	       );
 
 	# Main processing loop.
@@ -140,6 +157,10 @@ int _yp_parse(
 	// Parser input limits (Ragel internals).
 	const char *p, *pe, *eof;
 
+	// Current item indent.
+	size_t indent = 0;
+	// Current id dash position.
+	size_t id_pos = 0;
 	// Indicates if the current parsing step contains an item.
 	bool found = false;
 

tests/yparser.c

@@ -22,53 +22,56 @@
 #include "libknot/internal/yparser/yparser.h"
 #include "libknot/errcode.h"
 
-const char *syntax =
+const char *syntax_ok =
 	"a:\n"
 	"a :\n"
 	"a : #comment\n"
 	"\n"
-	"b:\"b\"\n"
 	"b: \"b\"\n"
-	"b:b#comment\n"
 	"b: b #comment\n"
-	"b :b\n"
 	"b : b\n"
-	"b: [b]#comment\n"
 	"b: [ b] # comment\n"
 	"b: [b ]\n"
 	"b: [ b ]\n"
 	"\n"
-	" f:\"f\"\n"
 	" f: \"f\"\n"
-	" f:f#comment\n"
 	" f: f #comment\n"
-	" f :f\n"
 	" f : f\n"
-	" f: [f]#comment\n"
 	" f: [ f] # comment\n"
 	" f: [f ]\n"
 	" f: [ f ]\n"
 	"\n"
-	"c: a b\n"
-	"c: a,b\n"
-	"c: a, b\n"
-	"c: a ,b\n"
-	"c: a , b\n"
-	"c: [a b]\n"
 	"c: [a,b]\n"
 	"c: [a, b]\n"
 	"c: [a ,b]\n"
 	"c: [a , b]\n"
 	"c: [ a , b ]\n"
 	"\n"
-	"-d:d\n"
-	" -d:d\n"
-	" - d:d\n"
-	" - d: d\n"
-	" - d :d\n"
-	" - d : d\n"
+	"- d: d\n"
+	"- d : d # comment\n"
 	"\n"
-	"e: \"a#b' c[d,]\"";
+	"e: \"a#b' c[d,]\"\n"
+	"\n"
+	"zone:\n"
+	"  -   domain: example.\n"
+	"      master: bind\n"
+	"  - domain: example.\n"
+	"    master: bind\n";
+
+const char *syntax_error1 =
+	"f:\n"
+	"  -  a: b\n"
+	"   - b: c\n";
+
+const char *syntax_error2 =
+	"f:\n"
+	"  -  a: b\n"
+	"      c: d\n";
+
+const char *syntax_error3 =
+	"f:\n"
+	"   a: b\n"
+	"  c: d\n";
 
 int main(int argc, char *argv[])
 {
@@ -81,7 +84,8 @@ int main(int argc, char *argv[])
 	yp_parser_t *yp = &yparser;
 	yp_init(yp);
 
-	ret = yp_set_input_string(yp, syntax, strlen(syntax));
+	// OK input.
+	ret = yp_set_input_string(yp, syntax_ok, strlen(syntax_ok));
 	ok(ret == KNOT_EOK, "set input string");
 
 	line = 1;
@@ -94,7 +98,7 @@ int main(int argc, char *argv[])
 	}
 
 	line = 5;
-	for (int i = 0; i < 10; i++) {
+	for (int i = 0; i < 6; i++) {
 		ret = yp_parse(yp);
 		ok(ret == KNOT_EOK, "parse %i. key0 with value", i);
 		ok(yp->key_len == 1 && yp->key[0] == 'b' &&
@@ -103,8 +107,8 @@ int main(int argc, char *argv[])
 		   "compare %i. key0 with value", i);
 	}
 
-	line = 16;
-	for (int i = 0; i < 10; i++) {
+	line = 12;
+	for (int i = 0; i < 6; i++) {
 		ret = yp_parse(yp);
 		ok(ret == KNOT_EOK, "parse %i. key1 with value", i);
 		ok(yp->key_len == 1 && yp->key[0] == 'f' &&
@@ -113,8 +117,8 @@ int main(int argc, char *argv[])
 		   "compare %i. key1 with value", i);
 	}
 
-	line = 27;
-	for (int i = 0; i < 11; i++) {
+	line = 19;
+	for (int i = 0; i < 5; i++) {
 		ret = yp_parse(yp);
 		ok(ret == KNOT_EOK, "parse %i. key0 with first value", i);
 		ok(yp->key_len == 1 && yp->key[0] == 'c' &&
@@ -130,8 +134,8 @@ int main(int argc, char *argv[])
 		   "compare %i. key0 with second value", i);
 	}
 
-	line = 39;
-	for (int i = 0; i < 6; i++) {
+	line = 25;
+	for (int i = 0; i < 2; i++) {
 		ret = yp_parse(yp);
 		ok(ret == KNOT_EOK, "parse %i. id", i);
 		ok(yp->key_len == 1 && yp->key[0] == 'd' &&
@@ -140,7 +144,7 @@ int main(int argc, char *argv[])
 		   "compare %i. id", i);
 	}
 
-	line = 46;
+	line = 28;
 	ret = yp_parse(yp);
 	ok(ret == KNOT_EOK, "parse key0 with quoted value");
 	ok(yp->key_len == 1 && yp->key[0] == 'e' && yp->data_len == 10 &&
@@ -148,9 +152,63 @@ int main(int argc, char *argv[])
 	   yp->event == YP_EKEY0 && yp->line_count == line,
 	   "compare key0 with quoted value");
 
+	line = 30;
+	ret = yp_parse(yp);
+	ok(ret == KNOT_EOK, "parse key0");
+	ok(yp->key_len == 4 && strcmp(yp->key, "zone") == 0 &&
+	   yp->data_len == 0 &&
+	   yp->event == YP_EKEY0 && yp->line_count == line,
+	   "compare key0 value");
+
+	line = 31;
+	for (int i = 0; i < 2; i++) {
+		ret = yp_parse(yp);
+		ok(ret == KNOT_EOK, "parse %i. id", i);
+		ok(yp->key_len == 6 && strcmp(yp->key, "domain") == 0 &&
+		   yp->data_len == 8 && strcmp(yp->data, "example.") == 0 &&
+		   yp->event == YP_EID && yp->line_count == line + 2 * i,
+		   "compare id");
+		ret = yp_parse(yp);
+		ok(ret == KNOT_EOK, "parse %i. key1", i);
+		ok(yp->key_len == 6 && strcmp(yp->key, "master") == 0 &&
+		   yp->data_len == 4 && strcmp(yp->data, "bind") == 0 &&
+		   yp->event == YP_EKEY1 && yp->line_count == line + 2 * i + 1,
+		   "compare key1");
+	}