WolodiaM
3 周前
提交者
GitHub
GitHub
找不到此签名对应的密钥
GPG 密钥 ID: B5690EEEBB952194
共有 3 个文件被更改，包括 365 次插入 和 4 次删除
分列视图
Diff 选项
-
+25 -4grammar.js
-
+148 -0src/scanner.c
-
+192 -0test/corpus/expressions.txt
+ 25
- 4
grammar.js
查看文件
+ 148
- 0
src/scanner.c
查看文件
| @ -0,0 +1,148 @@ | |||
| #include "tree_sitter/alloc.h" | |||
| #include "tree_sitter/parser.h" | |||
| #include <assert.h> | |||
| #include <string.h> | |||
| #include <wctype.h> | |||
| enum TokenType { GNU_RAW_STRING_DELIMITER, GNU_RAW_STRING_CONTENT }; | |||
| /// The spec limits delimiters to 16 chars | |||
| #define MAX_DELIMITER_LENGTH 16 | |||
| typedef struct { | |||
| uint8_t delimiter_length; | |||
| wchar_t delimiter[MAX_DELIMITER_LENGTH]; | |||
| } Scanner; | |||
| static inline void advance(TSLexer *lexer) { lexer->advance(lexer, false); } | |||
| static inline void reset(Scanner *scanner) { | |||
| scanner->delimiter_length = 0; | |||
| memset(scanner->delimiter, 0, sizeof scanner->delimiter); | |||
| } | |||
| /// Scan the raw string delimiter in R"delimiter(content)delimiter" | |||
| static bool scan_gnu_raw_string_delimiter(Scanner *scanner, TSLexer *lexer) { | |||
| if (scanner->delimiter_length > 0) { | |||
| // Closing delimiter: must exactly match the opening delimiter. | |||
| // We already checked this when scanning content, but this is how we | |||
| // know when to stop. We can't stop at ", because R"""hello""" is valid. | |||
| for (int i = 0; i < scanner->delimiter_length; ++i) { | |||
| if (lexer->lookahead != scanner->delimiter[i]) { | |||
| return false; | |||
| } | |||
| advance(lexer); | |||
| } | |||
| reset(scanner); | |||
| return true; | |||
| } | |||
| // Opening delimiter: record the d-char-sequence up to (. | |||
| // d-char is any basic character except parens, backslashes, and spaces. | |||
| for (;;) { | |||
| if (scanner->delimiter_length >= MAX_DELIMITER_LENGTH || lexer->eof(lexer) || lexer->lookahead == '\\' || | |||
| iswspace(lexer->lookahead)) { | |||
| return false; | |||
| } | |||
| if (lexer->lookahead == '(') { | |||
| // Rather than create a token for an empty delimiter, we fail and | |||
| // let the grammar fall back to a delimiter-less rule. | |||
| return scanner->delimiter_length > 0; | |||
| } | |||
| scanner->delimiter[scanner->delimiter_length++] = lexer->lookahead; | |||
| advance(lexer); | |||
| } | |||
| } | |||
| /// Scan the raw string content in R"delimiter(content)delimiter" | |||
| static bool scan_gnu_raw_string_content(Scanner *scanner, TSLexer *lexer) { | |||
| // The progress made through the delimiter since the last ')'. | |||
| // The delimiter may not contain ')' so a single counter suffices. | |||
| for (int delimiter_index = -1;;) { | |||
| // If we hit EOF, consider the content to terminate there. | |||
| // This forms an incomplete gnu_raw_string_literal, and models the code | |||
| // well. | |||
| if (lexer->eof(lexer)) { | |||
| lexer->mark_end(lexer); | |||
| return true; | |||
| } | |||
| if (delimiter_index >= 0) { | |||
| if (delimiter_index == scanner->delimiter_length) { | |||
| if (lexer->lookahead == '"') { | |||
| return true; | |||
| } | |||
| delimiter_index = -1; | |||
| } else { | |||
| if (lexer->lookahead == scanner->delimiter[delimiter_index]) { | |||
| delimiter_index += 1; | |||
| } else { | |||
| delimiter_index = -1; | |||
| } | |||
| } | |||
| } | |||
| if (delimiter_index == -1 && lexer->lookahead == ')') { | |||
| // The content doesn't include the )delimiter" part. | |||
| // We must still scan through it, but exclude it from the token. | |||
| lexer->mark_end(lexer); | |||
| delimiter_index = 0; | |||
| } | |||
| advance(lexer); | |||
| } | |||
| } | |||
| void *tree_sitter_c_external_scanner_create() { | |||
| Scanner *scanner = (Scanner *)ts_calloc(1, sizeof(Scanner)); | |||
| memset(scanner, 0, sizeof(Scanner)); | |||
| return scanner; | |||
| } | |||
| bool tree_sitter_c_external_scanner_scan(void *payload, TSLexer *lexer, const bool *valid_symbols) { | |||
| Scanner *scanner = (Scanner *)payload; | |||
| if (valid_symbols[GNU_RAW_STRING_DELIMITER] && valid_symbols[GNU_RAW_STRING_CONTENT]) { | |||
| // we're in error recovery | |||
| return false; | |||
| } | |||
| // No skipping leading whitespace: raw-string grammar is space-sensitive. | |||
| if (valid_symbols[GNU_RAW_STRING_DELIMITER]) { | |||
| lexer->result_symbol = GNU_RAW_STRING_DELIMITER; | |||
| return scan_gnu_raw_string_delimiter(scanner, lexer); | |||
| } | |||
| if (valid_symbols[GNU_RAW_STRING_CONTENT]) { | |||
| lexer->result_symbol = GNU_RAW_STRING_CONTENT; | |||
| return scan_gnu_raw_string_content(scanner, lexer); | |||
| } | |||
| return false; | |||
| } | |||
| unsigned tree_sitter_c_external_scanner_serialize(void *payload, char *buffer) { | |||
| static_assert(MAX_DELIMITER_LENGTH * sizeof(wchar_t) < TREE_SITTER_SERIALIZATION_BUFFER_SIZE, | |||
| "Serialized delimiter is too long!"); | |||
| Scanner *scanner = (Scanner *)payload; | |||
| size_t size = scanner->delimiter_length * sizeof(wchar_t); | |||
| memcpy(buffer, scanner->delimiter, size); | |||
| return (unsigned)size; | |||
| } | |||
| void tree_sitter_c_external_scanner_deserialize(void *payload, const char *buffer, unsigned length) { | |||
| assert(length % sizeof(wchar_t) == 0 && "Can't decode serialized delimiter!"); | |||
| Scanner *scanner = (Scanner *)payload; | |||
| scanner->delimiter_length = length / sizeof(wchar_t); | |||
| if (length > 0) { | |||
| memcpy(&scanner->delimiter[0], buffer, length); | |||
| } | |||
| } | |||
| void tree_sitter_c_external_scanner_destroy(void *payload) { | |||
| Scanner *scanner = (Scanner *)payload; | |||
| ts_free(scanner); | |||
| } | |||