/* +----------------------------------------------------------------------+ | Zend Engine | +----------------------------------------------------------------------+ | Copyright (c) 1998-2014 Zend Technologies Ltd. (http://www.zend.com) | +----------------------------------------------------------------------+ | This source file is subject to version 2.00 of the Zend license, | | that is bundled with this package in the file LICENSE, and is | | available through the world-wide-web at the following url: | | http://www.zend.com/license/2_00.txt. | | If you did not receive a copy of the Zend license and are unable to | | obtain it through the world-wide-web, please send a note to | | license@zend.com so we can mail you a copy immediately. | +----------------------------------------------------------------------+ | Authors: Marcus Boerger | | Nuno Lopes | | Scott MacVicar | | Flex version authors: | | Andi Gutmans | | Zeev Suraski | +----------------------------------------------------------------------+ */ /* $Id$ */ #if 0 # define YYDEBUG(s, c) printf("state: %d char: %c\n", s, c) #else # define YYDEBUG(s, c) #endif #include "zend_language_scanner_defs.h" #include #include "zend.h" #ifdef PHP_WIN32 # include #endif #include "zend_alloc.h" #include #include "zend_compile.h" #include "zend_language_scanner.h" #include "zend_highlight.h" #include "zend_constants.h" #include "zend_variables.h" #include "zend_operators.h" #include "zend_API.h" #include "zend_strtod.h" #include "zend_exceptions.h" #include "zend_virtual_cwd.h" #include "tsrm_config_common.h" #define YYCTYPE unsigned char #define YYFILL(n) { if ((YYCURSOR + n) >= (YYLIMIT + ZEND_MMAP_AHEAD)) { return 0; } } #define YYCURSOR SCNG(yy_cursor) #define YYLIMIT SCNG(yy_limit) #define YYMARKER SCNG(yy_marker) #define YYGETCONDITION() SCNG(yy_state) #define YYSETCONDITION(s) SCNG(yy_state) = s #define STATE(name) yyc##name /* emulate flex constructs */ #define BEGIN(state) YYSETCONDITION(STATE(state)) #define YYSTATE YYGETCONDITION() #define yytext ((char*)SCNG(yy_text)) #define yyleng SCNG(yy_leng) #define yyless(x) do { YYCURSOR = (unsigned char*)yytext + x; \ yyleng = (unsigned int)x; } while(0) #define yymore() goto yymore_restart /* perform sanity check. If this message is triggered you should increase the ZEND_MMAP_AHEAD value in the zend_streams.h file */ /*!max:re2c */ #if ZEND_MMAP_AHEAD < YYMAXFILL # error ZEND_MMAP_AHEAD should be greater than or equal to YYMAXFILL #endif #ifdef HAVE_STDARG_H # include #endif #ifdef HAVE_UNISTD_H # include #endif /* Globals Macros */ #define SCNG LANG_SCNG #ifdef ZTS ZEND_API ts_rsrc_id language_scanner_globals_id; #else ZEND_API zend_php_scanner_globals language_scanner_globals; #endif #define HANDLE_NEWLINES(s, l) \ do { \ char *p = (s), *boundary = p+(l); \ \ while (p= 'a' && (c) <= 'z') || ((c) >= 'A' && (c) <= 'Z') || (c) == '_' || (c) >= 0x7F) #define ZEND_IS_OCT(c) ((c)>='0' && (c)<='7') #define ZEND_IS_HEX(c) (((c)>='0' && (c)<='9') || ((c)>='a' && (c)<='f') || ((c)>='A' && (c)<='F')) BEGIN_EXTERN_C() static size_t encoding_filter_script_to_internal(unsigned char **to, size_t *to_length, const unsigned char *from, size_t from_length TSRMLS_DC) { const zend_encoding *internal_encoding = zend_multibyte_get_internal_encoding(TSRMLS_C); assert(internal_encoding && zend_multibyte_check_lexer_compatibility(internal_encoding)); return zend_multibyte_encoding_converter(to, to_length, from, from_length, internal_encoding, LANG_SCNG(script_encoding) TSRMLS_CC); } static size_t encoding_filter_script_to_intermediate(unsigned char **to, size_t *to_length, const unsigned char *from, size_t from_length TSRMLS_DC) { return zend_multibyte_encoding_converter(to, to_length, from, from_length, zend_multibyte_encoding_utf8, LANG_SCNG(script_encoding) TSRMLS_CC); } static size_t encoding_filter_intermediate_to_script(unsigned char **to, size_t *to_length, const unsigned char *from, size_t from_length TSRMLS_DC) { return zend_multibyte_encoding_converter(to, to_length, from, from_length, LANG_SCNG(script_encoding), zend_multibyte_encoding_utf8 TSRMLS_CC); } static size_t encoding_filter_intermediate_to_internal(unsigned char **to, size_t *to_length, const unsigned char *from, size_t from_length TSRMLS_DC) { const zend_encoding *internal_encoding = zend_multibyte_get_internal_encoding(TSRMLS_C); assert(internal_encoding && zend_multibyte_check_lexer_compatibility(internal_encoding)); return zend_multibyte_encoding_converter(to, to_length, from, from_length, internal_encoding, zend_multibyte_encoding_utf8 TSRMLS_CC); } static void _yy_push_state(int new_state TSRMLS_DC) { zend_stack_push(&SCNG(state_stack), (void *) &YYGETCONDITION(), sizeof(int)); YYSETCONDITION(new_state); } #define yy_push_state(state_and_tsrm) _yy_push_state(yyc##state_and_tsrm) static void yy_pop_state(TSRMLS_D) { int *stack_state; zend_stack_top(&SCNG(state_stack), (void **) &stack_state); YYSETCONDITION(*stack_state); zend_stack_del_top(&SCNG(state_stack)); } static void yy_scan_buffer(char *str, unsigned int len TSRMLS_DC) { YYCURSOR = (YYCTYPE*)str; YYLIMIT = YYCURSOR + len; if (!SCNG(yy_start)) { SCNG(yy_start) = YYCURSOR; } } void startup_scanner(TSRMLS_D) { CG(parse_error) = 0; CG(doc_comment) = NULL; zend_stack_init(&SCNG(state_stack)); zend_ptr_stack_init(&SCNG(heredoc_label_stack)); } static void heredoc_label_dtor(zend_heredoc_label *heredoc_label) { efree(heredoc_label->label); } void shutdown_scanner(TSRMLS_D) { CG(parse_error) = 0; RESET_DOC_COMMENT(); zend_stack_destroy(&SCNG(state_stack)); zend_ptr_stack_clean(&SCNG(heredoc_label_stack), (void (*)(void *)) &heredoc_label_dtor, 1); zend_ptr_stack_destroy(&SCNG(heredoc_label_stack)); } ZEND_API void zend_save_lexical_state(zend_lex_state *lex_state TSRMLS_DC) { lex_state->yy_leng = SCNG(yy_leng); lex_state->yy_start = SCNG(yy_start); lex_state->yy_text = SCNG(yy_text); lex_state->yy_cursor = SCNG(yy_cursor); lex_state->yy_marker = SCNG(yy_marker); lex_state->yy_limit = SCNG(yy_limit); lex_state->state_stack = SCNG(state_stack); zend_stack_init(&SCNG(state_stack)); lex_state->heredoc_label_stack = SCNG(heredoc_label_stack); zend_ptr_stack_init(&SCNG(heredoc_label_stack)); lex_state->in = SCNG(yy_in); lex_state->yy_state = YYSTATE; lex_state->filename = zend_get_compiled_filename(TSRMLS_C); lex_state->lineno = CG(zend_lineno); lex_state->script_org = SCNG(script_org); lex_state->script_org_size = SCNG(script_org_size); lex_state->script_filtered = SCNG(script_filtered); lex_state->script_filtered_size = SCNG(script_filtered_size); lex_state->input_filter = SCNG(input_filter); lex_state->output_filter = SCNG(output_filter); lex_state->script_encoding = SCNG(script_encoding); } ZEND_API void zend_restore_lexical_state(zend_lex_state *lex_state TSRMLS_DC) { SCNG(yy_leng) = lex_state->yy_leng; SCNG(yy_start) = lex_state->yy_start; SCNG(yy_text) = lex_state->yy_text; SCNG(yy_cursor) = lex_state->yy_cursor; SCNG(yy_marker) = lex_state->yy_marker; SCNG(yy_limit) = lex_state->yy_limit; zend_stack_destroy(&SCNG(state_stack)); SCNG(state_stack) = lex_state->state_stack; zend_ptr_stack_clean(&SCNG(heredoc_label_stack), (void (*)(void *)) &heredoc_label_dtor, 1); zend_ptr_stack_destroy(&SCNG(heredoc_label_stack)); SCNG(heredoc_label_stack) = lex_state->heredoc_label_stack; SCNG(yy_in) = lex_state->in; YYSETCONDITION(lex_state->yy_state); CG(zend_lineno) = lex_state->lineno; zend_restore_compiled_filename(lex_state->filename TSRMLS_CC); if (SCNG(script_filtered)) { efree(SCNG(script_filtered)); SCNG(script_filtered) = NULL; } SCNG(script_org) = lex_state->script_org; SCNG(script_org_size) = lex_state->script_org_size; SCNG(script_filtered) = lex_state->script_filtered; SCNG(script_filtered_size) = lex_state->script_filtered_size; SCNG(input_filter) = lex_state->input_filter; SCNG(output_filter) = lex_state->output_filter; SCNG(script_encoding) = lex_state->script_encoding; RESET_DOC_COMMENT(); } ZEND_API void zend_destroy_file_handle(zend_file_handle *file_handle TSRMLS_DC) { zend_llist_del_element(&CG(open_files), file_handle, (int (*)(void *, void *)) zend_compare_file_handles); /* zend_file_handle_dtor() operates on the copy, so we have to NULLify the original here */ file_handle->opened_path = NULL; if (file_handle->free_filename) { file_handle->filename = NULL; } } #define BOM_UTF32_BE "\x00\x00\xfe\xff" #define BOM_UTF32_LE "\xff\xfe\x00\x00" #define BOM_UTF16_BE "\xfe\xff" #define BOM_UTF16_LE "\xff\xfe" #define BOM_UTF8 "\xef\xbb\xbf" static const zend_encoding *zend_multibyte_detect_utf_encoding(const unsigned char *script, size_t script_size TSRMLS_DC) { const unsigned char *p; int wchar_size = 2; int le = 0; /* utf-16 or utf-32? */ p = script; while ((p-script) < script_size) { p = memchr(p, 0, script_size-(p-script)-2); if (!p) { break; } if (*(p+1) == '\0' && *(p+2) == '\0') { wchar_size = 4; break; } /* searching for UTF-32 specific byte orders, so this will do */ p += 4; } /* BE or LE? */ p = script; while ((p-script) < script_size) { if (*p == '\0' && *(p+wchar_size-1) != '\0') { /* BE */ le = 0; break; } else if (*p != '\0' && *(p+wchar_size-1) == '\0') { /* LE* */ le = 1; break; } p += wchar_size; } if (wchar_size == 2) { return le ? zend_multibyte_encoding_utf16le : zend_multibyte_encoding_utf16be; } else { return le ? zend_multibyte_encoding_utf32le : zend_multibyte_encoding_utf32be; } return NULL; } static const zend_encoding* zend_multibyte_detect_unicode(TSRMLS_D) { const zend_encoding *script_encoding = NULL; int bom_size; unsigned char *pos1, *pos2; if (LANG_SCNG(script_org_size) < sizeof(BOM_UTF32_LE)-1) { return NULL; } /* check out BOM */ if (!memcmp(LANG_SCNG(script_org), BOM_UTF32_BE, sizeof(BOM_UTF32_BE)-1)) { script_encoding = zend_multibyte_encoding_utf32be; bom_size = sizeof(BOM_UTF32_BE)-1; } else if (!memcmp(LANG_SCNG(script_org), BOM_UTF32_LE, sizeof(BOM_UTF32_LE)-1)) { script_encoding = zend_multibyte_encoding_utf32le; bom_size = sizeof(BOM_UTF32_LE)-1; } else if (!memcmp(LANG_SCNG(script_org), BOM_UTF16_BE, sizeof(BOM_UTF16_BE)-1)) { script_encoding = zend_multibyte_encoding_utf16be; bom_size = sizeof(BOM_UTF16_BE)-1; } else if (!memcmp(LANG_SCNG(script_org), BOM_UTF16_LE, sizeof(BOM_UTF16_LE)-1)) { script_encoding = zend_multibyte_encoding_utf16le; bom_size = sizeof(BOM_UTF16_LE)-1; } else if (!memcmp(LANG_SCNG(script_org), BOM_UTF8, sizeof(BOM_UTF8)-1)) { script_encoding = zend_multibyte_encoding_utf8; bom_size = sizeof(BOM_UTF8)-1; } if (script_encoding) { /* remove BOM */ LANG_SCNG(script_org) += bom_size; LANG_SCNG(script_org_size) -= bom_size; return script_encoding; } /* script contains NULL bytes -> auto-detection */ if ((pos1 = memchr(LANG_SCNG(script_org), 0, LANG_SCNG(script_org_size)))) { /* check if the NULL byte is after the __HALT_COMPILER(); */ pos2 = LANG_SCNG(script_org); while (pos1 - pos2 >= sizeof("__HALT_COMPILER();")-1) { pos2 = memchr(pos2, '_', pos1 - pos2); if (!pos2) break; pos2++; if (strncasecmp((char*)pos2, "_HALT_COMPILER", sizeof("_HALT_COMPILER")-1) == 0) { pos2 += sizeof("_HALT_COMPILER")-1; while (*pos2 == ' ' || *pos2 == '\t' || *pos2 == '\r' || *pos2 == '\n') { pos2++; } if (*pos2 == '(') { pos2++; while (*pos2 == ' ' || *pos2 == '\t' || *pos2 == '\r' || *pos2 == '\n') { pos2++; } if (*pos2 == ')') { pos2++; while (*pos2 == ' ' || *pos2 == '\t' || *pos2 == '\r' || *pos2 == '\n') { pos2++; } if (*pos2 == ';') { return NULL; } } } } } /* make best effort if BOM is missing */ return zend_multibyte_detect_utf_encoding(LANG_SCNG(script_org), LANG_SCNG(script_org_size) TSRMLS_CC); } return NULL; } static const zend_encoding* zend_multibyte_find_script_encoding(TSRMLS_D) { const zend_encoding *script_encoding; if (CG(detect_unicode)) { /* check out bom(byte order mark) and see if containing wchars */ script_encoding = zend_multibyte_detect_unicode(TSRMLS_C); if (script_encoding != NULL) { /* bom or wchar detection is prior to 'script_encoding' option */ return script_encoding; } } /* if no script_encoding specified, just leave alone */ if (!CG(script_encoding_list) || !CG(script_encoding_list_size)) { return NULL; } /* if multiple encodings specified, detect automagically */ if (CG(script_encoding_list_size) > 1) { return zend_multibyte_encoding_detector(LANG_SCNG(script_org), LANG_SCNG(script_org_size), CG(script_encoding_list), CG(script_encoding_list_size) TSRMLS_CC); } return CG(script_encoding_list)[0]; } ZEND_API int zend_multibyte_set_filter(const zend_encoding *onetime_encoding TSRMLS_DC) { const zend_encoding *internal_encoding = zend_multibyte_get_internal_encoding(TSRMLS_C); const zend_encoding *script_encoding = onetime_encoding ? onetime_encoding: zend_multibyte_find_script_encoding(TSRMLS_C); if (!script_encoding) { return FAILURE; } /* judge input/output filter */ LANG_SCNG(script_encoding) = script_encoding; LANG_SCNG(input_filter) = NULL; LANG_SCNG(output_filter) = NULL; if (!internal_encoding || LANG_SCNG(script_encoding) == internal_encoding) { if (!zend_multibyte_check_lexer_compatibility(LANG_SCNG(script_encoding))) { /* and if not, work around w/ script_encoding -> utf-8 -> script_encoding conversion */ LANG_SCNG(input_filter) = encoding_filter_script_to_intermediate; LANG_SCNG(output_filter) = encoding_filter_intermediate_to_script; } else { LANG_SCNG(input_filter) = NULL; LANG_SCNG(output_filter) = NULL; } return SUCCESS; } if (zend_multibyte_check_lexer_compatibility(internal_encoding)) { LANG_SCNG(input_filter) = encoding_filter_script_to_internal; LANG_SCNG(output_filter) = NULL; } else if (zend_multibyte_check_lexer_compatibility(LANG_SCNG(script_encoding))) { LANG_SCNG(input_filter) = NULL; LANG_SCNG(output_filter) = encoding_filter_script_to_internal; } else { /* both script and internal encodings are incompatible w/ flex */ LANG_SCNG(input_filter) = encoding_filter_script_to_intermediate; LANG_SCNG(output_filter) = encoding_filter_intermediate_to_internal; } return 0; } ZEND_API int open_file_for_scanning(zend_file_handle *file_handle TSRMLS_DC) { const char *file_path = NULL; char *buf; size_t size, offset = 0; zend_string *compiled_filename; /* The shebang line was read, get the current position to obtain the buffer start */ if (CG(start_lineno) == 2 && file_handle->type == ZEND_HANDLE_FP && file_handle->handle.fp) { if ((offset = ftell(file_handle->handle.fp)) == -1) { offset = 0; } } if (zend_stream_fixup(file_handle, &buf, &size TSRMLS_CC) == FAILURE) { return FAILURE; } zend_llist_add_element(&CG(open_files), file_handle); if (file_handle->handle.stream.handle >= (void*)file_handle && file_handle->handle.stream.handle <= (void*)(file_handle+1)) { zend_file_handle *fh = (zend_file_handle*)zend_llist_get_last(&CG(open_files)); size_t diff = (char*)file_handle->handle.stream.handle - (char*)file_handle; fh->handle.stream.handle = (void*)(((char*)fh) + diff); file_handle->handle.stream.handle = fh->handle.stream.handle; } /* Reset the scanner for scanning the new file */ SCNG(yy_in) = file_handle; SCNG(yy_start) = NULL; if (size != -1) { if (CG(multibyte)) { SCNG(script_org) = (unsigned char*)buf; SCNG(script_org_size) = size; SCNG(script_filtered) = NULL; zend_multibyte_set_filter(NULL TSRMLS_CC); if (SCNG(input_filter)) { if ((size_t)-1 == SCNG(input_filter)(&SCNG(script_filtered), &SCNG(script_filtered_size), SCNG(script_org), SCNG(script_org_size) TSRMLS_CC)) { zend_error_noreturn(E_COMPILE_ERROR, "Could not convert the script from the detected " "encoding \"%s\" to a compatible encoding", zend_multibyte_get_encoding_name(LANG_SCNG(script_encoding))); } buf = (char*)SCNG(script_filtered); size = SCNG(script_filtered_size); } } SCNG(yy_start) = (unsigned char *)buf - offset; yy_scan_buffer(buf, size TSRMLS_CC); } else { zend_error_noreturn(E_COMPILE_ERROR, "zend_stream_mmap() failed"); } BEGIN(INITIAL); if (file_handle->opened_path) { file_path = file_handle->opened_path; } else { file_path = file_handle->filename; } compiled_filename = STR_INIT(file_path, strlen(file_path), 0); zend_set_compiled_filename(compiled_filename TSRMLS_CC); STR_RELEASE(compiled_filename); if (CG(start_lineno)) { CG(zend_lineno) = CG(start_lineno); CG(start_lineno) = 0; } else { CG(zend_lineno) = 1; } RESET_DOC_COMMENT(); CG(increment_lineno) = 0; return SUCCESS; } END_EXTERN_C() ZEND_API zend_op_array *compile_file(zend_file_handle *file_handle, int type TSRMLS_DC) { zend_lex_state original_lex_state; zend_op_array *op_array = (zend_op_array *) emalloc(sizeof(zend_op_array)); zend_op_array *original_active_op_array = CG(active_op_array); zend_op_array *retval=NULL; int compiler_result; zend_bool compilation_successful=0; znode retval_znode; zend_bool original_in_compilation = CG(in_compilation); retval_znode.op_type = IS_CONST; ZVAL_LONG(&retval_znode.u.constant, 1); zend_save_lexical_state(&original_lex_state TSRMLS_CC); retval = op_array; /* success oriented */ if (open_file_for_scanning(file_handle TSRMLS_CC)==FAILURE) { if (type==ZEND_REQUIRE) { zend_message_dispatcher(ZMSG_FAILED_REQUIRE_FOPEN, file_handle->filename TSRMLS_CC); zend_bailout(); } else { zend_message_dispatcher(ZMSG_FAILED_INCLUDE_FOPEN, file_handle->filename TSRMLS_CC); } compilation_successful=0; } else { init_op_array(op_array, ZEND_USER_FUNCTION, INITIAL_OP_ARRAY_SIZE TSRMLS_CC); CG(in_compilation) = 1; CG(active_op_array) = op_array; zend_stack_push(&CG(context_stack), (void *) &CG(context), sizeof(CG(context))); zend_init_compiler_context(TSRMLS_C); compiler_result = zendparse(TSRMLS_C); zend_do_return(&retval_znode, 0 TSRMLS_CC); CG(in_compilation) = original_in_compilation; if (compiler_result != 0) { /* parser error */ zend_bailout(); } compilation_successful=1; } if (retval) { CG(active_op_array) = original_active_op_array; if (compilation_successful) { pass_two(op_array TSRMLS_CC); zend_release_labels(0 TSRMLS_CC); } else { efree(op_array); retval = NULL; } } zend_restore_lexical_state(&original_lex_state TSRMLS_CC); return retval; } zend_op_array *compile_filename(int type, zval *filename TSRMLS_DC) { zend_file_handle file_handle; zval tmp; zend_op_array *retval; char *opened_path = NULL; if (filename->type != IS_STRING) { tmp = *filename; zval_copy_ctor(&tmp); convert_to_string(&tmp); filename = &tmp; } file_handle.filename = Z_STRVAL_P(filename); file_handle.free_filename = 0; file_handle.type = ZEND_HANDLE_FILENAME; file_handle.opened_path = NULL; file_handle.handle.fp = NULL; retval = zend_compile_file(&file_handle, type TSRMLS_CC); if (retval && file_handle.handle.stream.handle) { if (!file_handle.opened_path) { file_handle.opened_path = opened_path = estrndup(Z_STRVAL_P(filename), Z_STRLEN_P(filename)); } zend_hash_str_add_empty_element(&EG(included_files), file_handle.opened_path, strlen(file_handle.opened_path)); if (opened_path) { efree(opened_path); } } zend_destroy_file_handle(&file_handle TSRMLS_CC); if (filename==&tmp) { zval_dtor(&tmp); } return retval; } ZEND_API int zend_prepare_string_for_scanning(zval *str, char *filename TSRMLS_DC) { char *buf; size_t size, old_len; zend_string *new_compiled_filename; /* enforce ZEND_MMAP_AHEAD trailing NULLs for flex... */ old_len = Z_STRLEN_P(str); Z_STR_P(str) = STR_REALLOC(Z_STR_P(str), old_len + ZEND_MMAP_AHEAD, 0); memset(Z_STRVAL_P(str) + old_len, 0, ZEND_MMAP_AHEAD + 1); SCNG(yy_in) = NULL; SCNG(yy_start) = NULL; buf = Z_STRVAL_P(str); size = old_len; if (CG(multibyte)) { SCNG(script_org) = (unsigned char*)buf; SCNG(script_org_size) = size; SCNG(script_filtered) = NULL; zend_multibyte_set_filter(zend_multibyte_get_internal_encoding(TSRMLS_C) TSRMLS_CC); if (SCNG(input_filter)) { if ((size_t)-1 == SCNG(input_filter)(&SCNG(script_filtered), &SCNG(script_filtered_size), SCNG(script_org), SCNG(script_org_size) TSRMLS_CC)) { zend_error_noreturn(E_COMPILE_ERROR, "Could not convert the script from the detected " "encoding \"%s\" to a compatible encoding", zend_multibyte_get_encoding_name(LANG_SCNG(script_encoding))); } buf = (char*)SCNG(script_filtered); size = SCNG(script_filtered_size); } } yy_scan_buffer(buf, size TSRMLS_CC); new_compiled_filename = STR_INIT(filename, strlen(filename), 0); zend_set_compiled_filename(new_compiled_filename TSRMLS_CC); STR_RELEASE(new_compiled_filename); CG(zend_lineno) = 1; CG(increment_lineno) = 0; RESET_DOC_COMMENT(); return SUCCESS; } ZEND_API size_t zend_get_scanned_file_offset(TSRMLS_D) { size_t offset = SCNG(yy_cursor) - SCNG(yy_start); if (SCNG(input_filter)) { size_t original_offset = offset, length = 0; do { unsigned char *p = NULL; if ((size_t)-1 == SCNG(input_filter)(&p, &length, SCNG(script_org), offset TSRMLS_CC)) { return (size_t)-1; } efree(p); if (length > original_offset) { offset--; } else if (length < original_offset) { offset++; } } while (original_offset != length); } return offset; } zend_op_array *compile_string(zval *source_string, char *filename TSRMLS_DC) { zend_lex_state original_lex_state; zend_op_array *op_array = (zend_op_array *) emalloc(sizeof(zend_op_array)); zend_op_array *original_active_op_array = CG(active_op_array); zend_op_array *retval; zval tmp; int compiler_result; zend_bool original_in_compilation = CG(in_compilation); if (Z_STRLEN_P(source_string)==0) { efree(op_array); return NULL; } CG(in_compilation) = 1; ZVAL_DUP(&tmp, source_string); convert_to_string(&tmp); source_string = &tmp; zend_save_lexical_state(&original_lex_state TSRMLS_CC); if (zend_prepare_string_for_scanning(source_string, filename TSRMLS_CC)==FAILURE) { efree(op_array); retval = NULL; } else { zend_bool orig_interactive = CG(interactive); CG(interactive) = 0; init_op_array(op_array, ZEND_EVAL_CODE, INITIAL_OP_ARRAY_SIZE TSRMLS_CC); CG(interactive) = orig_interactive; CG(active_op_array) = op_array; zend_stack_push(&CG(context_stack), (void *) &CG(context), sizeof(CG(context))); zend_init_compiler_context(TSRMLS_C); BEGIN(ST_IN_SCRIPTING); compiler_result = zendparse(TSRMLS_C); if (SCNG(script_filtered)) { efree(SCNG(script_filtered)); SCNG(script_filtered) = NULL; } if (compiler_result != 0) { CG(active_op_array) = original_active_op_array; CG(unclean_shutdown)=1; destroy_op_array(op_array TSRMLS_CC); efree(op_array); retval = NULL; } else { zend_do_return(NULL, 0 TSRMLS_CC); CG(active_op_array) = original_active_op_array; pass_two(op_array TSRMLS_CC); zend_release_labels(0 TSRMLS_CC); retval = op_array; } } zend_restore_lexical_state(&original_lex_state TSRMLS_CC); zval_dtor(&tmp); CG(in_compilation) = original_in_compilation; return retval; } BEGIN_EXTERN_C() int highlight_file(char *filename, zend_syntax_highlighter_ini *syntax_highlighter_ini TSRMLS_DC) { zend_lex_state original_lex_state; zend_file_handle file_handle; file_handle.type = ZEND_HANDLE_FILENAME; file_handle.filename = filename; file_handle.free_filename = 0; file_handle.opened_path = NULL; zend_save_lexical_state(&original_lex_state TSRMLS_CC); if (open_file_for_scanning(&file_handle TSRMLS_CC)==FAILURE) { zend_message_dispatcher(ZMSG_FAILED_HIGHLIGHT_FOPEN, filename TSRMLS_CC); zend_restore_lexical_state(&original_lex_state TSRMLS_CC); return FAILURE; } zend_highlight(syntax_highlighter_ini TSRMLS_CC); if (SCNG(script_filtered)) { efree(SCNG(script_filtered)); SCNG(script_filtered) = NULL; } zend_destroy_file_handle(&file_handle TSRMLS_CC); zend_restore_lexical_state(&original_lex_state TSRMLS_CC); return SUCCESS; } int highlight_string(zval *str, zend_syntax_highlighter_ini *syntax_highlighter_ini, char *str_name TSRMLS_DC) { zend_lex_state original_lex_state; zval tmp = *str; str = &tmp; zval_copy_ctor(str); zend_save_lexical_state(&original_lex_state TSRMLS_CC); if (zend_prepare_string_for_scanning(str, str_name TSRMLS_CC)==FAILURE) { zend_restore_lexical_state(&original_lex_state TSRMLS_CC); return FAILURE; } BEGIN(INITIAL); zend_highlight(syntax_highlighter_ini TSRMLS_CC); if (SCNG(script_filtered)) { efree(SCNG(script_filtered)); SCNG(script_filtered) = NULL; } zend_restore_lexical_state(&original_lex_state TSRMLS_CC); zval_dtor(str); return SUCCESS; } ZEND_API void zend_multibyte_yyinput_again(zend_encoding_filter old_input_filter, const zend_encoding *old_encoding TSRMLS_DC) { size_t length; unsigned char *new_yy_start; /* convert and set */ if (!SCNG(input_filter)) { if (SCNG(script_filtered)) { efree(SCNG(script_filtered)); SCNG(script_filtered) = NULL; } SCNG(script_filtered_size) = 0; length = SCNG(script_org_size); new_yy_start = SCNG(script_org); } else { if ((size_t)-1 == SCNG(input_filter)(&new_yy_start, &length, SCNG(script_org), SCNG(script_org_size) TSRMLS_CC)) { zend_error_noreturn(E_COMPILE_ERROR, "Could not convert the script from the detected " "encoding \"%s\" to a compatible encoding", zend_multibyte_get_encoding_name(LANG_SCNG(script_encoding))); } SCNG(script_filtered) = new_yy_start; SCNG(script_filtered_size) = length; } SCNG(yy_cursor) = new_yy_start + (SCNG(yy_cursor) - SCNG(yy_start)); SCNG(yy_marker) = new_yy_start + (SCNG(yy_marker) - SCNG(yy_start)); SCNG(yy_text) = new_yy_start + (SCNG(yy_text) - SCNG(yy_start)); SCNG(yy_limit) = new_yy_start + (SCNG(yy_limit) - SCNG(yy_start)); SCNG(yy_start) = new_yy_start; } # define zend_copy_value(zendlval, yytext, yyleng) \ if (SCNG(output_filter)) { \ size_t sz = 0; \ SCNG(output_filter)((unsigned char **)&Z_STRVAL_P(zendlval), &sz, (unsigned char *)yytext, (size_t)yyleng TSRMLS_CC); \ Z_STRLEN_P(zendlval) = sz; \ } else { \ ZVAL_STRINGL(zendlval, yytext, yyleng); \ } static void zend_scan_escape_string(zval *zendlval, char *str, int len, char quote_type TSRMLS_DC) { register char *s, *t; char *end; ZVAL_STRINGL(zendlval, str, len); /* convert escape sequences */ s = t = Z_STRVAL_P(zendlval); end = s+Z_STRLEN_P(zendlval); while (s= end) { *t++ = '\\'; break; } switch(*s) { case 'n': *t++ = '\n'; Z_STRLEN_P(zendlval)--; break; case 'r': *t++ = '\r'; Z_STRLEN_P(zendlval)--; break; case 't': *t++ = '\t'; Z_STRLEN_P(zendlval)--; break; case 'f': *t++ = '\f'; Z_STRLEN_P(zendlval)--; break; case 'v': *t++ = '\v'; Z_STRLEN_P(zendlval)--; break; case 'e': #ifdef PHP_WIN32 *t++ = VK_ESCAPE; #else *t++ = '\e'; #endif Z_STRLEN_P(zendlval)--; break; case '"': case '`': if (*s != quote_type) { *t++ = '\\'; *t++ = *s; break; } case '\\': case '$': *t++ = *s; Z_STRLEN_P(zendlval)--; break; case 'x': case 'X': if (ZEND_IS_HEX(*(s+1))) { char hex_buf[3] = { 0, 0, 0 }; Z_STRLEN_P(zendlval)--; /* for the 'x' */ hex_buf[0] = *(++s); Z_STRLEN_P(zendlval)--; if (ZEND_IS_HEX(*(s+1))) { hex_buf[1] = *(++s); Z_STRLEN_P(zendlval)--; } *t++ = (char) strtol(hex_buf, NULL, 16); } else { *t++ = '\\'; *t++ = *s; } break; default: /* check for an octal */ if (ZEND_IS_OCT(*s)) { char octal_buf[4] = { 0, 0, 0, 0 }; octal_buf[0] = *s; Z_STRLEN_P(zendlval)--; if (ZEND_IS_OCT(*(s+1))) { octal_buf[1] = *(++s); Z_STRLEN_P(zendlval)--; if (ZEND_IS_OCT(*(s+1))) { octal_buf[2] = *(++s); Z_STRLEN_P(zendlval)--; } } *t++ = (char) strtol(octal_buf, NULL, 8); } else { *t++ = '\\'; *t++ = *s; } break; } } else { *t++ = *s; } if (*s == '\n' || (*s == '\r' && (*(s+1) != '\n'))) { CG(zend_lineno)++; } s++; } *t = 0; if (SCNG(output_filter)) { size_t sz = 0; s = Z_STRVAL_P(zendlval); SCNG(output_filter)((unsigned char **)&Z_STRVAL_P(zendlval), &sz, (unsigned char *)s, (size_t)Z_STRLEN_P(zendlval) TSRMLS_CC); Z_STRLEN_P(zendlval) = sz; efree(s); } } int lex_scan(zval *zendlval TSRMLS_DC) { restart: SCNG(yy_text) = YYCURSOR; yymore_restart: /*!re2c re2c:yyfill:check = 0; LNUM [0-9]+ DNUM ([0-9]*"."[0-9]+)|([0-9]+"."[0-9]*) EXPONENT_DNUM (({LNUM}|{DNUM})[eE][+-]?{LNUM}) HNUM "0x"[0-9a-fA-F]+ BNUM "0b"[01]+ LABEL [a-zA-Z_\x7f-\xff][a-zA-Z0-9_\x7f-\xff]* WHITESPACE [ \n\r\t]+ TABS_AND_SPACES [ \t]* TOKENS [;:,.\[\]()|^&+-/*=%!~$<>?@] ANY_CHAR [^] NEWLINE ("\r"|"\n"|"\r\n") /* compute yyleng before each rule */ := yyleng = YYCURSOR - SCNG(yy_text); "exit" { return T_EXIT; } "die" { return T_EXIT; } "function" { return T_FUNCTION; } "const" { return T_CONST; } "return" { return T_RETURN; } "yield" { return T_YIELD; } "try" { return T_TRY; } "catch" { return T_CATCH; } "finally" { return T_FINALLY; } "throw" { return T_THROW; } "if" { return T_IF; } "elseif" { return T_ELSEIF; } "endif" { return T_ENDIF; } "else" { return T_ELSE; } "while" { return T_WHILE; } "endwhile" { return T_ENDWHILE; } "do" { return T_DO; } "for" { return T_FOR; } "endfor" { return T_ENDFOR; } "foreach" { return T_FOREACH; } "endforeach" { return T_ENDFOREACH; } "declare" { return T_DECLARE; } "enddeclare" { return T_ENDDECLARE; } "instanceof" { return T_INSTANCEOF; } "as" { return T_AS; } "switch" { return T_SWITCH; } "endswitch" { return T_ENDSWITCH; } "case" { return T_CASE; } "default" { return T_DEFAULT; } "break" { return T_BREAK; } "continue" { return T_CONTINUE; } "goto" { return T_GOTO; } "echo" { return T_ECHO; } "print" { return T_PRINT; } "class" { return T_CLASS; } "interface" { return T_INTERFACE; } "trait" { return T_TRAIT; } "extends" { return T_EXTENDS; } "implements" { return T_IMPLEMENTS; } "->" { yy_push_state(ST_LOOKING_FOR_PROPERTY TSRMLS_CC); return T_OBJECT_OPERATOR; } {WHITESPACE}+ { //??? ZVAL_STRINGL(zendlval, yytext, yyleng); ZVAL_STR(zendlval, STR_EMPTY_ALLOC()); /* tricky way to no copying - intentional */ HANDLE_NEWLINES(yytext, yyleng); return T_WHITESPACE; } "->" { return T_OBJECT_OPERATOR; } {LABEL} { yy_pop_state(TSRMLS_C); zend_copy_value(zendlval, yytext, yyleng); return T_STRING; } {ANY_CHAR} { yyless(0); yy_pop_state(TSRMLS_C); goto restart; } "::" { return T_PAAMAYIM_NEKUDOTAYIM; } "\\" { return T_NS_SEPARATOR; } "..." { return T_ELLIPSIS; } "new" { return T_NEW; } "clone" { return T_CLONE; } "var" { return T_VAR; } "("{TABS_AND_SPACES}("int"|"integer"){TABS_AND_SPACES}")" { return T_INT_CAST; } "("{TABS_AND_SPACES}("real"|"double"|"float"){TABS_AND_SPACES}")" { return T_DOUBLE_CAST; } "("{TABS_AND_SPACES}("string"|"binary"){TABS_AND_SPACES}")" { return T_STRING_CAST; } "("{TABS_AND_SPACES}"array"{TABS_AND_SPACES}")" { return T_ARRAY_CAST; } "("{TABS_AND_SPACES}"object"{TABS_AND_SPACES}")" { return T_OBJECT_CAST; } "("{TABS_AND_SPACES}("bool"|"boolean"){TABS_AND_SPACES}")" { return T_BOOL_CAST; } "("{TABS_AND_SPACES}("unset"){TABS_AND_SPACES}")" { return T_UNSET_CAST; } "eval" { return T_EVAL; } "include" { return T_INCLUDE; } "include_once" { return T_INCLUDE_ONCE; } "require" { return T_REQUIRE; } "require_once" { return T_REQUIRE_ONCE; } "namespace" { return T_NAMESPACE; } "use" { return T_USE; } "insteadof" { return T_INSTEADOF; } "global" { return T_GLOBAL; } "isset" { return T_ISSET; } "empty" { return T_EMPTY; } "__halt_compiler" { return T_HALT_COMPILER; } "static" { return T_STATIC; } "abstract" { return T_ABSTRACT; } "final" { return T_FINAL; } "private" { return T_PRIVATE; } "protected" { return T_PROTECTED; } "public" { return T_PUBLIC; } "unset" { return T_UNSET; } "=>" { return T_DOUBLE_ARROW; } "list" { return T_LIST; } "array" { return T_ARRAY; } "callable" { return T_CALLABLE; } "++" { return T_INC; } "--" { return T_DEC; } "===" { return T_IS_IDENTICAL; } "!==" { return T_IS_NOT_IDENTICAL; } "==" { return T_IS_EQUAL; } "!="|"<>" { return T_IS_NOT_EQUAL; } "<=" { return T_IS_SMALLER_OR_EQUAL; } ">=" { return T_IS_GREATER_OR_EQUAL; } "+=" { return T_PLUS_EQUAL; } "-=" { return T_MINUS_EQUAL; } "*=" { return T_MUL_EQUAL; } "/=" { return T_DIV_EQUAL; } ".=" { return T_CONCAT_EQUAL; } "%=" { return T_MOD_EQUAL; } "<<=" { return T_SL_EQUAL; } ">>=" { return T_SR_EQUAL; } "&=" { return T_AND_EQUAL; } "|=" { return T_OR_EQUAL; } "^=" { return T_XOR_EQUAL; } "||" { return T_BOOLEAN_OR; } "&&" { return T_BOOLEAN_AND; } "OR" { return T_LOGICAL_OR; } "AND" { return T_LOGICAL_AND; } "XOR" { return T_LOGICAL_XOR; } "<<" { return T_SL; } ">>" { return T_SR; } {TOKENS} { return yytext[0]; } "{" { yy_push_state(ST_IN_SCRIPTING TSRMLS_CC); return '{'; } "${" { yy_push_state(ST_LOOKING_FOR_VARNAME TSRMLS_CC); return T_DOLLAR_OPEN_CURLY_BRACES; } "}" { RESET_DOC_COMMENT(); if (!zend_stack_is_empty(&SCNG(state_stack))) { yy_pop_state(TSRMLS_C); } return '}'; } {LABEL}[[}] { yyless(yyleng - 1); zend_copy_value(zendlval, yytext, yyleng); yy_pop_state(TSRMLS_C); yy_push_state(ST_IN_SCRIPTING TSRMLS_CC); return T_STRING_VARNAME; } {ANY_CHAR} { yyless(0); yy_pop_state(TSRMLS_C); yy_push_state(ST_IN_SCRIPTING TSRMLS_CC); goto restart; } {BNUM} { char *bin = yytext + 2; /* Skip "0b" */ int len = yyleng - 2; /* Skip any leading 0s */ while (*bin == '0') { ++bin; --len; } if (len < SIZEOF_LONG * 8) { if (len == 0) { Z_LVAL_P(zendlval) = 0; } else { Z_LVAL_P(zendlval) = strtol(bin, NULL, 2); } zendlval->type = IS_LONG; return T_LNUMBER; } else { ZVAL_DOUBLE(zendlval, zend_bin_strtod(bin, NULL)); return T_DNUMBER; } } {LNUM} { if (yyleng < MAX_LENGTH_OF_LONG - 1) { /* Won't overflow */ Z_LVAL_P(zendlval) = strtol(yytext, NULL, 0); } else { errno = 0; Z_LVAL_P(zendlval) = strtol(yytext, NULL, 0); if (errno == ERANGE) { /* Overflow */ if (yytext[0] == '0') { /* octal overflow */ Z_DVAL_P(zendlval) = zend_oct_strtod(yytext, NULL); } else { Z_DVAL_P(zendlval) = zend_strtod(yytext, NULL); } zendlval->type = IS_DOUBLE; return T_DNUMBER; } } zendlval->type = IS_LONG; return T_LNUMBER; } {HNUM} { char *hex = yytext + 2; /* Skip "0x" */ int len = yyleng - 2; /* Skip any leading 0s */ while (*hex == '0') { hex++; len--; } if (len < SIZEOF_LONG * 2 || (len == SIZEOF_LONG * 2 && *hex <= '7')) { if (len == 0) { Z_LVAL_P(zendlval) = 0; } else { Z_LVAL_P(zendlval) = strtol(hex, NULL, 16); } zendlval->type = IS_LONG; return T_LNUMBER; } else { ZVAL_DOUBLE(zendlval, zend_hex_strtod(hex, NULL)); return T_DNUMBER; } } [0]|([1-9][0-9]*) { /* Offset could be treated as a long */ if (yyleng < MAX_LENGTH_OF_LONG - 1 || (yyleng == MAX_LENGTH_OF_LONG - 1 && strcmp(yytext, long_min_digits) < 0)) { ZVAL_LONG(zendlval, strtol(yytext, NULL, 10)); } else { ZVAL_STRINGL(zendlval, yytext, yyleng); } return T_NUM_STRING; } {LNUM}|{HNUM}|{BNUM} { /* Offset must be treated as a string */ ZVAL_STRINGL(zendlval, yytext, yyleng); return T_NUM_STRING; } {DNUM}|{EXPONENT_DNUM} { ZVAL_DOUBLE(zendlval, zend_strtod(yytext, NULL)); return T_DNUMBER; } "__CLASS__" { zend_class_entry *ce = CG(active_class_entry); if (ce && ZEND_ACC_TRAIT == (ce->ce_flags & ZEND_ACC_TRAIT)) { /* We create a special __CLASS__ constant that is going to be resolved at run-time */ ZVAL_STRINGL(zendlval, "__CLASS__", sizeof("__CLASS__") - 1); Z_TYPE_P(zendlval) = IS_CONSTANT; } else { if (ce && ce->name) { ZVAL_STR(zendlval, ce->name); Z_ADDREF_P(zendlval); } else { ZVAL_EMPTY_STRING(zendlval); } } return T_CLASS_C; } "__TRAIT__" { zend_class_entry *ce = CG(active_class_entry); if (ce && ce->name && ZEND_ACC_TRAIT == (ce->ce_flags & ZEND_ACC_TRAIT)) { ZVAL_STR(zendlval, ce->name); Z_ADDREF_P(zendlval); } else { ZVAL_EMPTY_STRING(zendlval); } return T_TRAIT_C; } "__FUNCTION__" { zend_op_array *op_array = CG(active_op_array); if (op_array && op_array->function_name) { ZVAL_STR(zendlval, STR_COPY(op_array->function_name)); } else { ZVAL_EMPTY_STRING(zendlval); } return T_FUNC_C; } "__METHOD__" { if (CG(active_class_entry)) { int len = 2; if (CG(active_class_entry)->name) { len += CG(active_class_entry)->name->len; } if (CG(active_op_array) && CG(active_op_array)->function_name) { len += CG(active_op_array)->function_name->len; } ZVAL_STR(zendlval, STR_ALLOC(len, 0)); len = 0; if (CG(active_class_entry)->name) { memcpy(Z_STRVAL_P(zendlval), CG(active_class_entry)->name->val, CG(active_class_entry)->name->len); len += CG(active_class_entry)->name->len; } memcpy(Z_STRVAL_P(zendlval) + len, "::", sizeof("::")-1); len += sizeof("::")-1; if (CG(active_op_array) && CG(active_op_array)->function_name) { memcpy(Z_STRVAL_P(zendlval) + len, CG(active_op_array)->function_name->val, CG(active_op_array)->function_name->len); len += CG(active_op_array)->function_name->len; } Z_STRVAL_P(zendlval)[len] = 0; } else if (CG(active_op_array) && CG(active_op_array)->function_name) { ZVAL_STR(zendlval, STR_COPY(CG(active_op_array)->function_name)); } else { ZVAL_EMPTY_STRING(zendlval); } return T_METHOD_C; } "__LINE__" { ZVAL_LONG(zendlval, CG(zend_lineno)); return T_LINE; } "__FILE__" { zend_string *filename = zend_get_compiled_filename(TSRMLS_C); if (!filename) { filename = STR_EMPTY_ALLOC(); } ZVAL_STR(zendlval, filename); return T_FILE; } "__DIR__" { zend_string *filename = zend_get_compiled_filename(TSRMLS_C); zend_string *dirname; if (!filename) { filename = STR_EMPTY_ALLOC(); } dirname = STR_DUP(filename, 0); zend_dirname(dirname->val, dirname->len); if (strcmp(dirname->val, ".") == 0) { dirname = STR_REALLOC(dirname, MAXPATHLEN, 0); #if HAVE_GETCWD VCWD_GETCWD(dirname->val, MAXPATHLEN); #elif HAVE_GETWD VCWD_GETWD(dirname->val); #endif } dirname->len = strlen(dirname->val); ZVAL_STR(zendlval, dirname); return T_DIR; } "__NAMESPACE__" { if (Z_TYPE(CG(current_namespace)) != IS_UNDEF) { ZVAL_DUP(zendlval, &CG(current_namespace)); } else { ZVAL_EMPTY_STRING(zendlval); } return T_NS_C; } "" { YYCTYPE *bracket = (YYCTYPE*)zend_memrchr(yytext, '<', yyleng - (sizeof("script language=php>") - 1)); if (bracket != SCNG(yy_text)) { /* Handle previously scanned HTML, as possible