diff options
Diffstat (limited to 'Zend/zend_language_scanner.l')
| -rw-r--r-- | Zend/zend_language_scanner.l | 145 |
1 files changed, 41 insertions, 104 deletions
diff --git a/Zend/zend_language_scanner.l b/Zend/zend_language_scanner.l index cf898f3bb2..7ac4efe353 100644 --- a/Zend/zend_language_scanner.l +++ b/Zend/zend_language_scanner.l @@ -919,7 +919,7 @@ DOUBLE_QUOTES_CHARS ("{"*([^$"\\{]|("\\"{ANY_CHAR}))|{DOUBLE_QUOTES_LITERAL_DOLL BACKQUOTE_CHARS ("{"*([^$`\\{]|("\\"{ANY_CHAR}))|{BACKQUOTE_LITERAL_DOLLAR}) HEREDOC_CHARS ("{"*([^$\n\r\\{]|("\\"[^\n\r]))|{HEREDOC_LITERAL_DOLLAR}|({HEREDOC_NEWLINE}+({HEREDOC_NON_LABEL}|{HEREDOC_LABEL_NO_NEWLINE}))) -NOWDOC_CHARS ({NEWLINE}*(([^a-zA-Z_\x7f-\xff\n\r][^\n\r]*)|({LABEL}[^a-zA-Z0-9_\x7f-\xff;\n\r][^\n\r]*)|({LABEL}[;][^\n\r]+))) +NOWDOC_CHARS ([^\n\r]|{NEWLINE}+([^a-zA-Z_\x7f-\xff\n\r]|({LABEL}([^a-zA-Z0-9_\x7f-\xff;\n\r]|(";"[^\n\r]))))) /* compute yyleng before each rule */ <!*> := yyleng = YYCURSOR - SCNG(yy_text); @@ -1896,7 +1896,7 @@ inline_char_handler: } -<ST_IN_SCRIPTING>b?"<<<"{TABS_AND_SPACES}({LABEL}|["]{LABEL}["]){NEWLINE} { +<ST_IN_SCRIPTING>b?"<<<"{TABS_AND_SPACES}({LABEL}|([']{LABEL}['])|(["]{LABEL}["])){NEWLINE} { char *s; int bprefix = (yytext[0] != '<') ? 1 : 0; @@ -1908,12 +1908,35 @@ inline_char_handler: CG(heredoc_len)--; } - if (*s == '"') { + if (*s == '\'') { s++; CG(heredoc_len) -= 2; + + BEGIN(ST_NOWDOC); + } else { + if (*s == '"') { + s++; + CG(heredoc_len) -= 2; + } + + BEGIN(ST_HEREDOC); } + CG(heredoc) = estrndup(s, CG(heredoc_len)); - BEGIN(ST_START_HEREDOC); + + /* Check for ending label on the next line */ + if (CG(heredoc_len) < YYLIMIT - YYCURSOR && !memcmp(YYCURSOR, s, CG(heredoc_len))) { + char *end = YYCURSOR + CG(heredoc_len); + + if (*end == ';') { + end++; + } + + if (*end == '\n' || *end == '\r') { + BEGIN(ST_END_HEREDOC); + } + } + return T_START_HEREDOC; } @@ -1924,34 +1947,6 @@ inline_char_handler: } -<ST_START_HEREDOC>{ANY_CHAR} { - yyless(0); - BEGIN(ST_HEREDOC); - goto restart; -} - -<ST_START_HEREDOC>{LABEL}";"?[\n\r] { - int label_len = yyleng - 1; - - if (yytext[label_len-1]==';') { - label_len--; - } - - yyless(label_len); - - if (label_len==CG(heredoc_len) && !memcmp(yytext, CG(heredoc), label_len)) { - zendlval->value.str.val = CG(heredoc); - zendlval->value.str.len = label_len; - CG(heredoc)=NULL; - CG(heredoc_len)=0; - BEGIN(ST_IN_SCRIPTING); - return T_END_HEREDOC; - } else { - BEGIN(ST_HEREDOC); - yymore(); - } -} - /* Match everything up to and including a possible ending label, so if the label * doesn't match, it's kept with the rest of the string * @@ -1979,12 +1974,9 @@ inline_char_handler: len--; /* Windows newline */ } - /* Go back before last label char, to match in ST_END_HEREDOC state */ - yyless(yyleng - 2); - - /* Subtract the remaining label length. yyleng must include newline - * before label, for zend_highlight/strip, tokenizer, etc. */ - yyleng -= CG(heredoc_len) - 1; + /* Go back before label, to match in ST_END_HEREDOC state. yytext will include + * newline before label, for zend_highlight/strip, tokenizer, etc. */ + yyless(yyleng - CG(heredoc_len) - 1); /* 1 for newline after label */ CG(increment_lineno) = 1; /* For newline before label */ BEGIN(ST_END_HEREDOC); @@ -1999,8 +1991,11 @@ inline_char_handler: } <ST_END_HEREDOC>{ANY_CHAR} { - SCNG(yy_text) = (unsigned char *)(Z_STRVAL_P(zendlval) = CG(heredoc)); - SCNG(yy_leng) = Z_STRLEN_P(zendlval) = CG(heredoc_len); + YYCURSOR += CG(heredoc_len) - 1; + yyleng = CG(heredoc_len); + + Z_STRVAL_P(zendlval) = CG(heredoc); + Z_STRLEN_P(zendlval) = CG(heredoc_len); CG(heredoc) = NULL; CG(heredoc_len) = 0; BEGIN(ST_IN_SCRIPTING); @@ -2065,52 +2060,7 @@ inline_char_handler: } -/* BEGIN nowdoc */ - -<ST_IN_SCRIPTING>b?"<<<"{TABS_AND_SPACES}[']{LABEL}[']{NEWLINE} { - int bprefix = (yytext[0] != '<') ? 1 : 0; - char *s; - CG(zend_lineno)++; - /* 3 is <<<, 2 is quotes, 1 is newline */ - CG(heredoc_len) = yyleng-bprefix-3-2-1-(yytext[yyleng-2]=='\r'?1:0); - s = yytext+bprefix+3; - while ((*s == ' ') || (*s == '\t')) { - s++; - CG(heredoc_len)--; - } - s++; /* first quote */ - CG(heredoc) = estrndup(s, CG(heredoc_len)); - BEGIN(ST_START_NOWDOC); - return T_START_NOWDOC; -} - -<ST_START_NOWDOC>{ANY_CHAR} { - yyless(0); - BEGIN(ST_NOWDOC); - goto restart; -} - -<ST_START_NOWDOC>{LABEL}";"?[\r\n] { - int label_len = yyleng - 1; - - if (yytext[label_len-1]==';') { - label_len--; - } - - if (label_len==CG(heredoc_len) && !memcmp(yytext, CG(heredoc), label_len)) { - yyless(label_len-1); - yyleng = 0; - BEGIN(ST_END_NOWDOC); - ZVAL_EMPTY_STRING(zendlval); - return T_ENCAPSED_AND_WHITESPACE; - } else { - BEGIN(ST_NOWDOC); - yyless(label_len); - yymore(); - } -} - -<ST_NOWDOC>{NOWDOC_CHARS}*{NEWLINE}+{LABEL}";"?[\n\r] { +<ST_NOWDOC>({NOWDOC_CHARS}+{NEWLINE}+|{NEWLINE}+){LABEL}";"?[\n\r] { char *end = yytext + yyleng - 1; if (end[-1] == ';') { @@ -2131,19 +2081,16 @@ inline_char_handler: len--; /* Windows newline */ } - /* Go back before last label char, to match in ST_END_NOWDOC state */ - yyless(yyleng - 2); - - /* Subtract the remaining label length. yyleng must include newline - * before label, for zend_highlight/strip, tokenizer, etc. */ - yyleng -= CG(heredoc_len) - 1; + /* Go back before label, to match in ST_END_HEREDOC state. yytext will include + * newline before label, for zend_highlight/strip, tokenizer, etc. */ + yyless(yyleng - CG(heredoc_len) - 1); /* 1 for newline after label */ CG(increment_lineno) = 1; /* For newline before label */ - BEGIN(ST_END_NOWDOC); + BEGIN(ST_END_HEREDOC); - HANDLE_NEWLINES(yytext, len); zend_copy_value(zendlval, yytext, len); zendlval->type = IS_STRING; + HANDLE_NEWLINES(yytext, len); return T_ENCAPSED_AND_WHITESPACE; } else { /* Go back to end of label, so the next match works correctly in case of @@ -2153,16 +2100,6 @@ inline_char_handler: } } -<ST_END_NOWDOC>{ANY_CHAR} { - SCNG(yy_text) = (unsigned char*)(Z_STRVAL_P(zendlval) = CG(heredoc)); - SCNG(yy_leng) = Z_STRLEN_P(zendlval) = CG(heredoc_len); - CG(heredoc) = NULL; - CG(heredoc_len) = 0; - BEGIN(ST_IN_SCRIPTING); - return T_END_NOWDOC; -} - -/* END nowdoc */ <ST_DOUBLE_QUOTES>["] { BEGIN(ST_IN_SCRIPTING); |