diff options
Diffstat (limited to 'json-glib/json-parser.c')
| -rw-r--r-- | json-glib/json-parser.c | 488 |
1 files changed, 200 insertions, 288 deletions
diff --git a/json-glib/json-parser.c b/json-glib/json-parser.c index 8df897e..3e00cb6 100644 --- a/json-glib/json-parser.c +++ b/json-glib/json-parser.c @@ -106,12 +106,12 @@ static guint parser_signals[LAST_SIGNAL] = { 0, }; G_DEFINE_TYPE (JsonParser, json_parser, G_TYPE_OBJECT); -static guint json_parse_array (JsonParser *parser, - JsonScanner *scanner, - gboolean nested); -static guint json_parse_object (JsonParser *parser, - JsonScanner *scanner, - gboolean nested); +static guint json_parse_array (JsonParser *parser, + JsonScanner *scanner, + JsonNode **node); +static guint json_parse_object (JsonParser *parser, + JsonScanner *scanner, + JsonNode **node); static inline void json_parser_clear (JsonParser *parser) @@ -423,385 +423,293 @@ json_parse_value (JsonParser *parser, } static guint -json_parse_array (JsonParser *parser, - JsonScanner *scanner, - gboolean nested) +json_parse_array (JsonParser *parser, + JsonScanner *scanner, + JsonNode **node) { JsonParserPrivate *priv = parser->priv; + JsonNode *old_current; JsonArray *array; guint token; + gint idx; - if (!nested) - { - /* the caller already swallowed the opening '[' */ - token = json_scanner_get_next_token (scanner); - if (token != G_TOKEN_LEFT_BRACE) - return G_TOKEN_LEFT_BRACE; - } - - g_signal_emit (parser, parser_signals[ARRAY_START], 0); + old_current = priv->current_node; + priv->current_node = json_node_new (JSON_NODE_ARRAY); array = json_array_new (); token = json_scanner_get_next_token (scanner); + g_assert (token == G_TOKEN_LEFT_BRACE); + + g_signal_emit (parser, parser_signals[ARRAY_START], 0); + + idx = 0; while (token != G_TOKEN_RIGHT_BRACE) { - JsonNode *node = NULL; + guint next_token = json_scanner_peek_next_token (scanner); + JsonNode *element = NULL; - /* nested object */ - if (token == G_TOKEN_LEFT_CURLY) + /* parse the element */ + switch (next_token) { - JsonNode *old_node = priv->current_node; - - priv->current_node = json_node_new (JSON_NODE_OBJECT); - - token = json_parse_object (parser, scanner, TRUE); - - node = priv->current_node; - priv->current_node = old_node; - - if (token != G_TOKEN_NONE) - { - json_node_free (node); - json_array_unref (array); - - return token; - } - - json_array_add_element (array, node); - json_node_set_parent (node, priv->current_node); - - g_signal_emit (parser, parser_signals[ARRAY_ELEMENT], 0, - array, - json_array_get_length (array)); - + case G_TOKEN_LEFT_BRACE: + JSON_NOTE (PARSER, "Nested array at index %d", idx); + token = json_parse_array (parser, scanner, &element); + break; + + case G_TOKEN_LEFT_CURLY: + JSON_NOTE (PARSER, "Nested object at index %d", idx); + token = json_parse_object (parser, scanner, &element); + break; + + case G_TOKEN_INT: + case G_TOKEN_FLOAT: + case G_TOKEN_STRING: + case '-': + case JSON_TOKEN_TRUE: + case JSON_TOKEN_FALSE: + case JSON_TOKEN_NULL: token = json_scanner_get_next_token (scanner); - if (token == G_TOKEN_RIGHT_BRACE) - { - break; - } - - if (token == G_TOKEN_COMMA) - { - token = json_scanner_get_next_token (scanner); + token = json_parse_value (parser, scanner, token, &element); + break; - if (token == G_TOKEN_RIGHT_BRACE) - { - json_array_unref (array); - priv->error_code = JSON_PARSER_ERROR_TRAILING_COMMA; - return G_TOKEN_SYMBOL; - } + case G_TOKEN_RIGHT_BRACE: + goto array_done; - continue; - } - - json_array_unref (array); - - return G_TOKEN_RIGHT_BRACE; + default: + if (next_token != G_TOKEN_RIGHT_BRACE) + token = G_TOKEN_RIGHT_BRACE; + break; } - /* nested array */ - if (token == G_TOKEN_LEFT_BRACE) + if (token != G_TOKEN_NONE || element == NULL) { - JsonNode *old_node = priv->current_node; - - priv->current_node = json_node_new (JSON_NODE_ARRAY); - - token = json_parse_array (parser, scanner, TRUE); - - node = priv->current_node; - priv->current_node = old_node; - - if (token != G_TOKEN_NONE) - { - json_node_free (node); - json_array_unref (array); - - return token; - } + /* the json_parse_* functions will have set the error code */ + json_array_unref (array); + json_node_free (priv->current_node); + priv->current_node = old_current; - json_array_add_element (array, node); - json_node_set_parent (node, priv->current_node); + return token; + } - g_signal_emit (parser, parser_signals[ARRAY_ELEMENT], 0, - array, - json_array_get_length (array)); + next_token = json_scanner_peek_next_token (scanner); + if (next_token == G_TOKEN_COMMA) + { token = json_scanner_get_next_token (scanner); - if (token == G_TOKEN_RIGHT_BRACE) - { - break; - } + next_token = json_scanner_peek_next_token (scanner); - if (token == G_TOKEN_COMMA) + /* look for trailing commas */ + if (next_token == G_TOKEN_RIGHT_BRACE) { - token = json_scanner_get_next_token (scanner); + priv->error_code = JSON_PARSER_ERROR_TRAILING_COMMA; - if (token == G_TOKEN_RIGHT_BRACE) - { - json_array_unref (array); - priv->error_code = JSON_PARSER_ERROR_TRAILING_COMMA; - return G_TOKEN_SYMBOL; - } + json_array_unref (array); + json_node_free (priv->current_node); + json_node_free (element); + priv->current_node = old_current; - continue; + return G_TOKEN_RIGHT_BRACE; } - - json_array_unref (array); - - return G_TOKEN_RIGHT_BRACE; - } - - /* value */ - token = json_parse_value (parser, scanner, token, &node); - if (token != G_TOKEN_NONE || node == NULL) - { - json_array_unref (array); - return token; } - json_array_add_element (array, node); - json_node_set_parent (node, priv->current_node); + JSON_NOTE (PARSER, "Array element %d completed", idx++); + json_node_set_parent (element, priv->current_node); + json_array_add_element (array, element); g_signal_emit (parser, parser_signals[ARRAY_ELEMENT], 0, array, - json_array_get_length (array)); - - token = json_scanner_get_next_token (scanner); - - if (token == G_TOKEN_COMMA) - { - token = json_scanner_get_next_token (scanner); + idx); - if (token == G_TOKEN_RIGHT_BRACE) - { - json_array_unref (array); - priv->error_code = JSON_PARSER_ERROR_TRAILING_COMMA; - return G_TOKEN_SYMBOL; - } - } - else if (token != G_TOKEN_RIGHT_BRACE) - { - json_array_unref (array); - priv->error_code = JSON_PARSER_ERROR_MISSING_COMMA; - return G_TOKEN_RIGHT_BRACE; - } + token = next_token; } +array_done: + json_scanner_get_next_token (scanner); + json_node_take_array (priv->current_node, array); + json_node_set_parent (priv->current_node, old_current); g_signal_emit (parser, parser_signals[ARRAY_END], 0, array); + if (node != NULL && *node == NULL) + *node = priv->current_node; + + priv->current_node = old_current; + return G_TOKEN_NONE; } static guint -json_parse_object (JsonParser *parser, - JsonScanner *scanner, - gboolean nested) +json_parse_object (JsonParser *parser, + JsonScanner *scanner, + JsonNode **node) { JsonParserPrivate *priv = parser->priv; JsonObject *object; + JsonNode *old_current; guint token; - if (!nested) - { - /* the caller already swallowed the opening '{' */ - token = json_scanner_get_next_token (scanner); - if (token != G_TOKEN_LEFT_CURLY) - return G_TOKEN_LEFT_CURLY; - } - - g_signal_emit (parser, parser_signals[OBJECT_START], 0); + old_current = priv->current_node; + priv->current_node = json_node_new (JSON_NODE_OBJECT); object = json_object_new (); token = json_scanner_get_next_token (scanner); + g_assert (token == G_TOKEN_LEFT_CURLY); + + g_signal_emit (parser, parser_signals[OBJECT_START], 0); + while (token != G_TOKEN_RIGHT_CURLY) { - JsonNode *node = NULL; - gchar *name = NULL; + guint next_token = json_scanner_peek_next_token (scanner); + JsonNode *member = NULL; + gchar *name; + + /* we need to abort here because empty objects do not + * have member names + */ + if (next_token == G_TOKEN_RIGHT_CURLY) + break; - if (token == G_TOKEN_STRING) + /* parse the member's name */ + if (next_token != G_TOKEN_STRING) { - name = g_strdup (scanner->value.v_string); + JSON_NOTE (PARSER, "Missing object member name"); - token = json_scanner_get_next_token (scanner); - if (token != ':') - { - g_free (name); - json_object_unref (object); + priv->error_code = JSON_PARSER_ERROR_PARSE; - return ':'; - } - else - { - /* swallow the colon */ - token = json_scanner_get_next_token (scanner); - } - } - - if (!name) - { json_object_unref (object); + json_node_free (priv->current_node); + priv->current_node = old_current; return G_TOKEN_STRING; } - /* nested object */ - if (token == G_TOKEN_LEFT_CURLY) - { - JsonNode *old_node = priv->current_node; - - priv->current_node = json_node_new (JSON_NODE_OBJECT); - - token = json_parse_object (parser, scanner, TRUE); - - node = priv->current_node; - priv->current_node = old_node; - - if (token != G_TOKEN_NONE) - { - g_free (name); - - if (node) - json_node_free (node); - - json_object_unref (object); - - return token; - } + /* member name */ + token = json_scanner_get_next_token (scanner); + name = g_strdup (scanner->value.v_string); + JSON_NOTE (PARSER, "Object member '%s'", name); - json_object_set_member (object, name, node); - json_node_set_parent (node, priv->current_node); + /* a colon separates names from values */ + next_token = json_scanner_peek_next_token (scanner); + if (next_token != ':') + { + JSON_NOTE (PARSER, "Missing object member name separator"); - g_signal_emit (parser, parser_signals[OBJECT_MEMBER], 0, - object, - name); + /* FIXME - MISSING_COLON */ + priv->error_code = JSON_PARSER_ERROR_PARSE; g_free (name); + json_object_unref (object); + json_node_free (priv->current_node); + priv->current_node = old_current; - token = json_scanner_get_next_token (scanner); - if (token == G_TOKEN_RIGHT_CURLY) - break; + return ':'; + } - if (token == G_TOKEN_COMMA) - { - token = json_scanner_get_next_token (scanner); + /* we swallow the ':' */ + token = json_scanner_get_next_token (scanner); + g_assert (token == ':'); + next_token = json_scanner_peek_next_token (scanner); - if (token == G_TOKEN_RIGHT_CURLY) - { - json_object_unref (object); - priv->error_code = JSON_PARSER_ERROR_TRAILING_COMMA; - return G_TOKEN_STRING; - } + /* parse the member's value */ + switch (next_token) + { + case G_TOKEN_LEFT_BRACE: + JSON_NOTE (PARSER, "Nested array at member %s", name); + token = json_parse_array (parser, scanner, &member); + break; + + case G_TOKEN_LEFT_CURLY: + JSON_NOTE (PARSER, "Nested object at member %s", name); + token = json_parse_object (parser, scanner, &member); + break; + + case G_TOKEN_INT: + case G_TOKEN_FLOAT: + case G_TOKEN_STRING: + case '-': + case JSON_TOKEN_TRUE: + case JSON_TOKEN_FALSE: + case JSON_TOKEN_NULL: + token = json_scanner_get_next_token (scanner); + token = json_parse_value (parser, scanner, token, &member); + break; - continue; - } + default: + /* once a member name is defined we need a value */ + token = G_TOKEN_SYMBOL; + break; + } + if (token != G_TOKEN_NONE || member == NULL) + { + /* the json_parse_* functions will have set the error code */ + g_free (name); json_object_unref (object); + json_node_free (priv->current_node); + priv->current_node = old_current; - return G_TOKEN_RIGHT_CURLY; + return token; } - /* nested array */ - if (token == G_TOKEN_LEFT_BRACE) + next_token = json_scanner_peek_next_token (scanner); + if (next_token == G_TOKEN_COMMA) { - JsonNode *old_node = priv->current_node; - - priv->current_node = json_node_new (JSON_NODE_ARRAY); - - token = json_parse_array (parser, scanner, TRUE); - - node = priv->current_node; - priv->current_node = old_node; - - if (token != G_TOKEN_NONE) - { - g_free (name); - json_node_free (node); - json_object_unref (object); - - return token; - } - - json_object_set_member (object, name, node); - json_node_set_parent (node, priv->current_node); - - g_signal_emit (parser, parser_signals[OBJECT_MEMBER], 0, - object, - name); - - g_free (name); - token = json_scanner_get_next_token (scanner); - if (token == G_TOKEN_RIGHT_CURLY) - break; + next_token = json_scanner_peek_next_token (scanner); - if (token == G_TOKEN_COMMA) + /* look for trailing commas */ + if (next_token == G_TOKEN_RIGHT_CURLY) { - token = json_scanner_get_next_token (scanner); + priv->error_code = JSON_PARSER_ERROR_TRAILING_COMMA; - if (token == G_TOKEN_RIGHT_CURLY) - { - json_object_unref (object); - priv->error_code = JSON_PARSER_ERROR_TRAILING_COMMA; - return G_TOKEN_STRING; - } + json_object_unref (object); + json_node_free (member); + json_node_free (priv->current_node); + priv->current_node = old_current; - continue; + return G_TOKEN_RIGHT_BRACE; } - - json_object_unref (object); - - return G_TOKEN_RIGHT_CURLY; } - - /* value */ - token = json_parse_value (parser, scanner, token, &node); - if (token != G_TOKEN_NONE || node == NULL) + else if (next_token == G_TOKEN_STRING) { + priv->error_code = JSON_PARSER_ERROR_MISSING_COMMA; + json_object_unref (object); - g_free (name); - return token; + json_node_free (member); + json_node_free (priv->current_node); + priv->current_node = old_current; + + return G_TOKEN_COMMA; } - json_object_set_member (object, name, node); - json_node_set_parent (node, priv->current_node); + JSON_NOTE (PARSER, "Object member '%s' completed", name); + json_node_set_parent (member, priv->current_node); + json_object_set_member (object, name, member); g_signal_emit (parser, parser_signals[OBJECT_MEMBER], 0, object, name); - token = json_scanner_get_next_token (scanner); - - if (token == G_TOKEN_COMMA) - { - token = json_scanner_get_next_token (scanner); - - if (token == G_TOKEN_RIGHT_CURLY) - { - g_free (name); - json_object_unref (object); - priv->error_code = JSON_PARSER_ERROR_TRAILING_COMMA; - return G_TOKEN_STRING; - } - } - else if (token != G_TOKEN_RIGHT_CURLY) - { - g_free (name); - json_object_unref (object); - priv->error_code = JSON_PARSER_ERROR_MISSING_COMMA; - return G_TOKEN_RIGHT_CURLY; - } - g_free (name); + + token = next_token; } + json_scanner_get_next_token (scanner); + json_node_take_object (priv->current_node, object); + json_node_set_parent (priv->current_node, old_current); g_signal_emit (parser, parser_signals[OBJECT_END], 0, object); + if (node != NULL && *node == NULL) + *node = priv->current_node; + + priv->current_node = old_current; + return G_TOKEN_NONE; } @@ -816,12 +724,12 @@ json_parse_statement (JsonParser *parser, switch (token) { case G_TOKEN_LEFT_CURLY: - priv->root = priv->current_node = json_node_new (JSON_NODE_OBJECT); - return json_parse_object (parser, scanner, FALSE); + JSON_NOTE (PARSER, "Statement is object declaration"); + return json_parse_object (parser, scanner, &priv->root); case G_TOKEN_LEFT_BRACE: - priv->root = priv->current_node = json_node_new (JSON_NODE_ARRAY); - return json_parse_array (parser, scanner, FALSE); + JSON_NOTE (PARSER, "Statement is array declaration"); + return json_parse_array (parser, scanner, &priv->root); /* some web APIs are not only passing the data structures: they are * also passing an assigment, which makes parsing horribly complicated @@ -833,6 +741,8 @@ json_parse_statement (JsonParser *parser, guint next_token; gchar *name; + JSON_NOTE (PARSER, "Statement is an assignment"); + /* swallow the 'var' token... */ token = json_scanner_get_next_token (scanner); @@ -875,10 +785,12 @@ json_parse_statement (JsonParser *parser, case G_TOKEN_INT: case G_TOKEN_FLOAT: case G_TOKEN_STRING: + JSON_NOTE (PARSER, "Statement is a value"); token = json_scanner_get_next_token (scanner); return json_parse_value (parser, scanner, token, &priv->root); default: + JSON_NOTE (PARSER, "Unknown statement"); json_scanner_get_next_token (scanner); priv->error_code = JSON_PARSER_ERROR_INVALID_BAREWORD; return G_TOKEN_SYMBOL; |