maride/php-parser
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
a488f43496
php-parser/scanner/scanner.l

659 lines
27 KiB
Plaintext
Raw Normal View History

split scanner and parser
2018-01-24 16:42:23 +00:00
%{
// Copyright (c) 2011 CZ.NIC z.s.p.o. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// blame: jnml, labs.nic.cz
package scanner
import (
"fmt"
"bytes"
issue #2 - fix template string scanning `l.Prev.Rune` is actually current rune and `c` is next rune.
2018-03-29 13:46:21 +00:00
"github.com/cznic/golex/lex"
split scanner and parser
2018-01-24 16:42:23 +00:00
)
const (
INITIAL = iota
PHP
STRING
STRING_VAR
STRING_VAR_INDEX
STRING_VAR_NAME
PROPERTY
HEREDOC_END
NOWDOC
HEREDOC
BACKQUOTE
)
var heredocLabel []lex.Char
issue #2 - fix template string scanning now it correctly check first char of variable
2018-03-30 09:17:25 +00:00
func isValidFirstVarNameRune(r rune) bool {
return r >= 'A' && r <= 'Z' || r == '_' || r >= 'a' && r <= 'z' || r >= '\u007f' && r <= 'ÿ'
}
split scanner and parser
2018-01-24 16:42:23 +00:00
func (l *Lexer) Lex(lval Lval) int {
l.Comments = nil
c := l.Enter()
%}
%s PHP STRING STRING_VAR STRING_VAR_INDEX STRING_VAR_NAME PROPERTY HEREDOC_END NOWDOC HEREDOC BACKQUOTE
%yyb last == '\n' || last = '\0'
%yyt l.getCurrentState()
%yyc c
%yyn c = l.Next()
%yym l.Mark()
%optioncase-insensitive
LNUM [0-9]+
DNUM ([0-9]*"."[0-9]+)|([0-9]+"."[0-9]*)
HNUM 0x[0-9a-fA-F]+
BNUM 0b[01]+
EXPONENT_DNUM (({LNUM}|{DNUM})[eE][+-]?{LNUM})
issue #2 - deny numbers and slashes in first char in variable name
2018-03-30 11:28:50 +00:00
VAR_NAME [a-zA-Z_\x7f-\xff][a-zA-Z0-9_\x7f-\xff]*
split scanner and parser
2018-01-24 16:42:23 +00:00
OPERATORS [;:,.\[\]()|\/\^&\+-*=%!~$<>?@]
NEW_LINE (\r|\n|\r\n)
%%
c = l.Rule0()
scanner: remove unnecessary lval.Token calls
2018-05-26 21:55:26 +00:00
<INITIAL>[ \t\n\r]+
split scanner and parser
2018-01-24 16:42:23 +00:00
<INITIAL>.
tb := []lex.Char{}
for {
if c == -1 {
tb = l.Token();
break;
}
if '?' == rune(c) {
tb = l.Token();
if (len(tb) < 2 || tb[len(tb)-1].Rune != '<') {
c = l.Next()
continue;
}
tb = l.ungetChars(1)
break;
}
c = l.Next()
}
lval.Token(l.newToken(tb))
return T_INLINE_HTML
scanner: remove unnecessary lval.Token calls
2018-05-26 21:55:26 +00:00
<INITIAL>\<\?php([ \t]|{NEW_LINE}) l.begin(PHP);
<INITIAL>\<\? l.begin(PHP);
split scanner and parser
2018-01-24 16:42:23 +00:00
<INITIAL>\<\?= l.begin(PHP);lval.Token(l.newToken(l.Token())); return T_ECHO;
scanner: remove unnecessary lval.Token calls
2018-05-26 21:55:26 +00:00
<PHP>[ \t\n\r]+
remove redundant nop before inline html
2018-04-01 21:02:13 +00:00
<PHP>[;][ \t\n\r]*\?\>{NEW_LINE}? l.begin(INITIAL);lval.Token(l.newToken(l.Token())); return Rune2Class(';');
split scanner and parser
2018-01-24 16:42:23 +00:00
<PHP>\?\>{NEW_LINE}? l.begin(INITIAL);lval.Token(l.newToken(l.Token())); return Rune2Class(';');
<PHP>{DNUM}|{EXPONENT_DNUM} lval.Token(l.newToken(l.Token())); return T_DNUMBER
<PHP>{BNUM}
tb := l.Token()
i:=2
BNUMFOR:for {
if i > len(tb)-1 {
break BNUMFOR;
}
switch tb[i].Rune {
case '0': i++;
default: break BNUMFOR;
}
}
if len(tb) - i < 64 {
lval.Token(l.newToken(l.Token())); return T_LNUMBER
} else {
lval.Token(l.newToken(l.Token())); return T_DNUMBER
}
<PHP>{LNUM}
if len(l.Token()) < 20 {
lval.Token(l.newToken(l.Token())); return T_LNUMBER
} else {
lval.Token(l.newToken(l.Token())); return T_DNUMBER
}
<PHP>{HNUM}
tb := l.Token()
i:=2
HNUMFOR:for {
if i > len(tb)-1 {
break HNUMFOR;
}
switch tb[i].Rune {
case '0': i++;
default: break HNUMFOR;
}
}
length := len(tb) - i
if length < 16 || (length == 16 && tb[i].Rune <= '7') {
lval.Token(l.newToken(l.Token())); return T_LNUMBER
} else {
lval.Token(l.newToken(l.Token())); return T_DNUMBER
}
<PHP>abstract lval.Token(l.newToken(l.Token())); return T_ABSTRACT
<PHP>array lval.Token(l.newToken(l.Token())); return T_ARRAY
<PHP>as lval.Token(l.newToken(l.Token())); return T_AS
<PHP>break lval.Token(l.newToken(l.Token())); return T_BREAK
<PHP>callable lval.Token(l.newToken(l.Token())); return T_CALLABLE
<PHP>case lval.Token(l.newToken(l.Token())); return T_CASE
<PHP>catch lval.Token(l.newToken(l.Token())); return T_CATCH
<PHP>class lval.Token(l.newToken(l.Token())); return T_CLASS
<PHP>clone lval.Token(l.newToken(l.Token())); return T_CLONE
scanner test
2018-02-17 10:53:10 +00:00
<PHP>const lval.Token(l.newToken(l.Token())); return T_CONST
<PHP>continue lval.Token(l.newToken(l.Token())); return T_CONTINUE
<PHP>declare lval.Token(l.newToken(l.Token())); return T_DECLARE
<PHP>default lval.Token(l.newToken(l.Token())); return T_DEFAULT
<PHP>do lval.Token(l.newToken(l.Token())); return T_DO
<PHP>echo lval.Token(l.newToken(l.Token())); return T_ECHO
<PHP>else lval.Token(l.newToken(l.Token())); return T_ELSE
<PHP>elseif lval.Token(l.newToken(l.Token())); return T_ELSEIF
<PHP>empty lval.Token(l.newToken(l.Token())); return T_EMPTY
split scanner and parser
2018-01-24 16:42:23 +00:00
<PHP>enddeclare lval.Token(l.newToken(l.Token())); return T_ENDDECLARE
<PHP>endfor lval.Token(l.newToken(l.Token())); return T_ENDFOR
<PHP>endforeach lval.Token(l.newToken(l.Token())); return T_ENDFOREACH
<PHP>endif lval.Token(l.newToken(l.Token())); return T_ENDIF
<PHP>endswitch lval.Token(l.newToken(l.Token())); return T_ENDSWITCH
<PHP>endwhile lval.Token(l.newToken(l.Token())); return T_ENDWHILE
<PHP>eval lval.Token(l.newToken(l.Token())); return T_EVAL
<PHP>exit|die lval.Token(l.newToken(l.Token())); return T_EXIT
<PHP>extends lval.Token(l.newToken(l.Token())); return T_EXTENDS
<PHP>final lval.Token(l.newToken(l.Token())); return T_FINAL
<PHP>finally lval.Token(l.newToken(l.Token())); return T_FINALLY
<PHP>for lval.Token(l.newToken(l.Token())); return T_FOR
<PHP>foreach lval.Token(l.newToken(l.Token())); return T_FOREACH
<PHP>function|cfunction lval.Token(l.newToken(l.Token())); return T_FUNCTION
<PHP>global lval.Token(l.newToken(l.Token())); return T_GLOBAL
<PHP>goto lval.Token(l.newToken(l.Token())); return T_GOTO
<PHP>if lval.Token(l.newToken(l.Token())); return T_IF
<PHP>isset lval.Token(l.newToken(l.Token())); return T_ISSET
<PHP>implements lval.Token(l.newToken(l.Token())); return T_IMPLEMENTS
<PHP>instanceof lval.Token(l.newToken(l.Token())); return T_INSTANCEOF
<PHP>insteadof lval.Token(l.newToken(l.Token())); return T_INSTEADOF
<PHP>interface lval.Token(l.newToken(l.Token())); return T_INTERFACE
<PHP>list lval.Token(l.newToken(l.Token())); return T_LIST
<PHP>namespace lval.Token(l.newToken(l.Token())); return T_NAMESPACE
<PHP>private lval.Token(l.newToken(l.Token())); return T_PRIVATE
<PHP>public lval.Token(l.newToken(l.Token())); return T_PUBLIC
<PHP>print lval.Token(l.newToken(l.Token())); return T_PRINT
<PHP>protected lval.Token(l.newToken(l.Token())); return T_PROTECTED
<PHP>return lval.Token(l.newToken(l.Token())); return T_RETURN
<PHP>static lval.Token(l.newToken(l.Token())); return T_STATIC
<PHP>switch lval.Token(l.newToken(l.Token())); return T_SWITCH
<PHP>throw lval.Token(l.newToken(l.Token())); return T_THROW
<PHP>trait lval.Token(l.newToken(l.Token())); return T_TRAIT
<PHP>try lval.Token(l.newToken(l.Token())); return T_TRY
<PHP>unset lval.Token(l.newToken(l.Token())); return T_UNSET
<PHP>use lval.Token(l.newToken(l.Token())); return T_USE
<PHP>var lval.Token(l.newToken(l.Token())); return T_VAR
<PHP>while lval.Token(l.newToken(l.Token())); return T_WHILE
<PHP>yield[ \t\n\r]+from[^a-zA-Z0-9_\x80-\xff] lval.Token(l.newToken(l.Token())); return T_YIELD_FROM
<PHP>yield lval.Token(l.newToken(l.Token())); return T_YIELD
<PHP>include lval.Token(l.newToken(l.Token())); return T_INCLUDE
<PHP>include_once lval.Token(l.newToken(l.Token())); return T_INCLUDE_ONCE
<PHP>require lval.Token(l.newToken(l.Token())); return T_REQUIRE
<PHP>require_once lval.Token(l.newToken(l.Token())); return T_REQUIRE_ONCE
<PHP>__CLASS__ lval.Token(l.newToken(l.Token())); return T_CLASS_C
<PHP>__DIR__ lval.Token(l.newToken(l.Token())); return T_DIR
<PHP>__FILE__ lval.Token(l.newToken(l.Token())); return T_FILE
<PHP>__FUNCTION__ lval.Token(l.newToken(l.Token())); return T_FUNC_C
<PHP>__LINE__ lval.Token(l.newToken(l.Token())); return T_LINE
<PHP>__NAMESPACE__ lval.Token(l.newToken(l.Token())); return T_NS_C
<PHP>__METHOD__ lval.Token(l.newToken(l.Token())); return T_METHOD_C
<PHP>__TRAIT__ lval.Token(l.newToken(l.Token())); return T_TRAIT_C
<PHP>__halt_compiler lval.Token(l.newToken(l.Token())); return T_HALT_COMPILER
<PHP>\([ \t]*array[ \t]*\) lval.Token(l.newToken(l.Token())); return T_ARRAY_CAST
<PHP>\([ \t]*(bool|boolean)[ \t]*\) lval.Token(l.newToken(l.Token())); return T_BOOL_CAST
<PHP>\([ \t]*(real|double|float)[ \t]*\) lval.Token(l.newToken(l.Token())); return T_DOUBLE_CAST
<PHP>\([ \t]*(int|integer)[ \t]*\) lval.Token(l.newToken(l.Token())); return T_INT_CAST
<PHP>\([ \t]*object[ \t]*\) lval.Token(l.newToken(l.Token())); return T_OBJECT_CAST
<PHP>\([ \t]*string[ \t]*\) lval.Token(l.newToken(l.Token())); return T_STRING_CAST
<PHP>\([ \t]*unset[ \t]*\) lval.Token(l.newToken(l.Token())); return T_UNSET_CAST
<PHP>new lval.Token(l.newToken(l.Token())); return T_NEW
<PHP>and lval.Token(l.newToken(l.Token())); return T_LOGICAL_AND
<PHP>or lval.Token(l.newToken(l.Token())); return T_LOGICAL_OR
<PHP>xor lval.Token(l.newToken(l.Token())); return T_LOGICAL_XOR
<PHP>\\ lval.Token(l.newToken(l.Token())); return T_NS_SEPARATOR
scanner test
2018-02-17 10:53:10 +00:00
<PHP>\.\.\. lval.Token(l.newToken(l.Token())); return T_ELLIPSIS
<PHP>:: lval.Token(l.newToken(l.Token())); return T_PAAMAYIM_NEKUDOTAYIM // T_DOUBLE_COLON
split scanner and parser
2018-01-24 16:42:23 +00:00
<PHP>&& lval.Token(l.newToken(l.Token())); return T_BOOLEAN_AND
<PHP>\|\| lval.Token(l.newToken(l.Token())); return T_BOOLEAN_OR
<PHP>&= lval.Token(l.newToken(l.Token())); return T_AND_EQUAL
<PHP>\|= lval.Token(l.newToken(l.Token())); return T_OR_EQUAL
scanner test
2018-02-17 10:53:10 +00:00
<PHP>\.= lval.Token(l.newToken(l.Token())); return T_CONCAT_EQUAL
split scanner and parser
2018-01-24 16:42:23 +00:00
<PHP>\*= lval.Token(l.newToken(l.Token())); return T_MUL_EQUAL
<PHP>\*\*= lval.Token(l.newToken(l.Token())); return T_POW_EQUAL
scanner test
2018-02-17 10:53:10 +00:00
<PHP>[/]= lval.Token(l.newToken(l.Token())); return T_DIV_EQUAL
split scanner and parser
2018-01-24 16:42:23 +00:00
<PHP>\+= lval.Token(l.newToken(l.Token())); return T_PLUS_EQUAL
<PHP>-= lval.Token(l.newToken(l.Token())); return T_MINUS_EQUAL
<PHP>\^= lval.Token(l.newToken(l.Token())); return T_XOR_EQUAL
<PHP>%= lval.Token(l.newToken(l.Token())); return T_MOD_EQUAL
scanner test
2018-02-17 10:53:10 +00:00
<PHP>-- lval.Token(l.newToken(l.Token())); return T_DEC
split scanner and parser
2018-01-24 16:42:23 +00:00
<PHP>\+\+ lval.Token(l.newToken(l.Token())); return T_INC
scanner test
2018-02-17 10:53:10 +00:00
<PHP>=> lval.Token(l.newToken(l.Token())); return T_DOUBLE_ARROW
split scanner and parser
2018-01-24 16:42:23 +00:00
<PHP>\<=\> lval.Token(l.newToken(l.Token())); return T_SPACESHIP
<PHP>\!=|\<\> lval.Token(l.newToken(l.Token())); return T_IS_NOT_EQUAL
<PHP>\!== lval.Token(l.newToken(l.Token())); return T_IS_NOT_IDENTICAL
<PHP>== lval.Token(l.newToken(l.Token())); return T_IS_EQUAL
<PHP>=== lval.Token(l.newToken(l.Token())); return T_IS_IDENTICAL
<PHP>\<\<= lval.Token(l.newToken(l.Token())); return T_SL_EQUAL
<PHP>\>\>= lval.Token(l.newToken(l.Token())); return T_SR_EQUAL
<PHP>\>= lval.Token(l.newToken(l.Token())); return T_IS_GREATER_OR_EQUAL
<PHP>\<= lval.Token(l.newToken(l.Token())); return T_IS_SMALLER_OR_EQUAL
<PHP>\*\* lval.Token(l.newToken(l.Token())); return T_POW
<PHP>\<\< lval.Token(l.newToken(l.Token())); return T_SL
<PHP>\>\> lval.Token(l.newToken(l.Token())); return T_SR
<PHP>\?\? lval.Token(l.newToken(l.Token())); return T_COALESCE
issue #2 - fix single line comment fix handling end of file add handling php close tag
2018-03-30 14:15:26 +00:00
<PHP>(#|[/][/])
merge DocComment and PlainComment
2018-04-15 12:55:33 +00:00
tb := l.Token()
issue #2 - fix single line comment fix handling end of file add handling php close tag
2018-03-30 14:15:26 +00:00
for {
if c == -1 {
break
}
merge DocComment and PlainComment
2018-04-15 12:55:33 +00:00
tb = append(tb, l.Last)
issue #2 - fix single line comment fix handling end of file add handling php close tag
2018-03-30 14:15:26 +00:00
switch c {
case '\r':
c = l.Next()
if c == '\n' {
continue
}
case '\n':
case '?':
c = l.Next()
if c == '>' {
l.ungetChars(1)
tb = tb[:len(tb)-1]
break
}
continue
default:
c = l.Next()
continue
}
break;
}
merge DocComment and PlainComment
2018-04-15 12:55:33 +00:00
l.addComment(tb)
issue #2 - fix single line comment fix handling end of file add handling php close tag
2018-03-30 14:15:26 +00:00
split scanner and parser
2018-01-24 16:42:23 +00:00
<PHP>([/][*])|([/][*][*])
tb := l.Token()
is_doc_comment := false
if len(tb) > 2 {
is_doc_comment = true
l.PhpDocComment = ""
}
for {
if c == -1 {
break; // TODO: Unterminated comment starting line %d
}
scanner: fix scanning empty inline comment
2018-04-10 21:58:57 +00:00
if l.Prev.Rune == '*' && l.Last.Rune == '/' {
split scanner and parser
2018-01-24 16:42:23 +00:00
c = l.Next()
break;
}
scanner: fix scanning empty inline comment
2018-04-10 21:58:57 +00:00
c = l.Next()
split scanner and parser
2018-01-24 16:42:23 +00:00
}
if is_doc_comment {
l.PhpDocComment = string(l.TokenBytes(nil))
merge DocComment and PlainComment
2018-04-15 12:55:33 +00:00
l.addComment(l.Token())
split scanner and parser
2018-01-24 16:42:23 +00:00
} else {
merge DocComment and PlainComment
2018-04-15 12:55:33 +00:00
l.addComment(l.Token())
split scanner and parser
2018-01-24 16:42:23 +00:00
}
<PHP>{OPERATORS} lval.Token(l.newToken(l.Token())); return Rune2Class(rune(l.TokenBytes(nil)[0]))
<PHP>\{ l.pushState(PHP); lval.Token(l.newToken(l.Token())); return Rune2Class(rune(l.TokenBytes(nil)[0]))
<PHP>\} l.popState(); lval.Token(l.newToken(l.Token())); return Rune2Class(rune(l.TokenBytes(nil)[0])); l.PhpDocComment = ""
<PHP>\${VAR_NAME} lval.Token(l.newToken(l.Token())); return T_VARIABLE
<PHP>{VAR_NAME} lval.Token(l.newToken(l.Token())); return T_STRING
<PHP>-> l.begin(PROPERTY);lval.Token(l.newToken(l.Token())); return T_OBJECT_OPERATOR;
scanner: remove unnecessary lval.Token calls
2018-05-26 21:55:26 +00:00
<PROPERTY>[ \t\n\r]+
split scanner and parser
2018-01-24 16:42:23 +00:00
<PROPERTY>-> lval.Token(l.newToken(l.Token())); return T_OBJECT_OPERATOR;
<PROPERTY>{VAR_NAME} l.begin(PHP);lval.Token(l.newToken(l.Token())); return T_STRING;
<PROPERTY>. l.ungetChars(1);l.begin(PHP)
<PHP>[\']([^\\\']*([\\].)*)*[\'] lval.Token(l.newToken(l.Token())); return T_CONSTANT_ENCAPSED_STRING;
<PHP>` l.begin(BACKQUOTE); lval.Token(l.newToken(l.Token())); return Rune2Class(rune(l.TokenBytes(nil)[0]))
<BACKQUOTE>` l.begin(PHP); lval.Token(l.newToken(l.Token())); return Rune2Class(rune(l.TokenBytes(nil)[0]))
<PHP>[b]?\<\<\<[ \t]*({VAR_NAME}|([']{VAR_NAME}['])|(["]{VAR_NAME}["])){NEW_LINE}
tb := l.Token()
binPrefix := 0
if tb[0].Rune == 'b' {
binPrefix = 1
}
lblFirst := 3 + binPrefix
lblLast := len(tb)-2
if tb[lblLast].Rune == '\r' {
lblLast--
}
for {
if tb[lblFirst].Rune == ' ' || tb[lblFirst].Rune == '\t' {
lblFirst++
continue
}
break
}
add Heredoc node
2018-04-05 10:47:36 +00:00
heredocToken := make([]lex.Char, lblLast - lblFirst + 1)
copy(heredocToken, tb[lblFirst:lblLast+1])
split scanner and parser
2018-01-24 16:42:23 +00:00
switch tb[lblFirst].Rune {
case '\'' :
lblFirst++
lblLast--
l.begin(NOWDOC)
case '"' :
lblFirst++
lblLast--
l.begin(HEREDOC)
default:
l.begin(HEREDOC)
}
heredocLabel = make([]lex.Char, lblLast - lblFirst + 1)
copy(heredocLabel, tb[lblFirst:lblLast+1])
ungetCnt := len(heredocLabel)
searchLabelAhead := []lex.Char{}
for i := 0; i < len(heredocLabel); i++ {
if c == -1 {
break;
}
searchLabelAhead = append(searchLabelAhead, l.Lookahead())
c = l.Next()
}
if bytes.Equal(l.charsToBytes(heredocLabel), l.charsToBytes(searchLabelAhead)) && ';' == rune(c) {
ungetCnt++
c = l.Next()
if '\n' == rune(c) || '\r' == rune(c) {
l.begin(HEREDOC_END)
}
}
l.ungetChars(ungetCnt)
add Heredoc node
2018-04-05 10:47:36 +00:00
lval.Token(l.newToken(heredocToken));
split scanner and parser
2018-01-24 16:42:23 +00:00
return T_START_HEREDOC
<NOWDOC>.|[ \t\n\r]
searchLabel := []byte{}
tb := []lex.Char{}
for {
if c == -1 {
break;
}
if '\n' == rune(c) || '\r' == rune(c) {
if bytes.Equal(append(l.charsToBytes(heredocLabel), ';'), searchLabel) {
l.begin(HEREDOC_END)
tb = l.ungetChars(len(heredocLabel)+1)
break;
}
if bytes.Equal(l.charsToBytes(heredocLabel), searchLabel) {
l.begin(HEREDOC_END)
tb = l.ungetChars(len(heredocLabel))
break;
}
searchLabel = []byte{}
} else {
searchLabel = append(searchLabel, byte(rune(c)))
}
c = l.Next()
}
lval.Token(l.newToken(tb) )
return T_ENCAPSED_AND_WHITESPACE
<HEREDOC_END>{VAR_NAME}\; l.begin(PHP);lval.Token(l.newToken(l.ungetChars(1))); return T_END_HEREDOC
<HEREDOC_END>{VAR_NAME} l.begin(PHP);lval.Token(l.newToken(l.Token())); return T_END_HEREDOC
<PHP>[b]?[\"]
binPrefix := l.Token()[0].Rune == 'b'
beginString := func() int {
cnt := 1; if (binPrefix) {cnt = 2}
l.ungetChars(len(l.Token())-cnt)
chars := l.Token()[:cnt]
l.pushState(STRING)
lval.Token(l.newToken(chars)); return Rune2Class('"')
}
F:for {
if c == -1 {
break;
}
switch c {
case '"' :
c = l.Next();
lval.Token(l.newToken(l.Token())); return T_CONSTANT_ENCAPSED_STRING
break F;
case '$':
c = l.Next();
if rune(c) == '{' || c >= 'A' && c <= 'Z' || c == '_' || c >= 'a' && c <= 'z' || c >= '\u007f' && c <= 'ÿ' {
return beginString()
break F;
}
l.ungetChars(0)
case '{':
c = l.Next();
if rune(c) == '$' {
return beginString()
break F;
}
l.ungetChars(0)
case '\\':
c = l.Next();
}
c = l.Next()
}
<STRING>\" l.popState(); lval.Token(l.newToken(l.Token())); return Rune2Class(l.Token()[0].Rune)
<STRING,HEREDOC,BACKQUOTE>\{\$ lval.Token(l.newToken(l.ungetChars(1))); l.pushState(PHP); return T_CURLY_OPEN
<STRING,HEREDOC,BACKQUOTE>\$\{ l.pushState(STRING_VAR_NAME); lval.Token(l.newToken(l.Token())); return T_DOLLAR_OPEN_CURLY_BRACES
issue #2 - deny numbers and slashes in first char in variable name
2018-03-30 11:28:50 +00:00
<STRING,HEREDOC,BACKQUOTE>\${VAR_NAME} l.ungetChars(len(l.Token()));l.pushState(STRING_VAR)
split scanner and parser
2018-01-24 16:42:23 +00:00
<STRING>.|[ \t\n\r]
issue #2 - deny numbers and slashes in first char in variable name
2018-03-30 11:28:50 +00:00
currentChar := l.Prev
tb := []lex.Char{currentChar}
for {
switch currentChar.Rune {
split scanner and parser
2018-01-24 16:42:23 +00:00
case '$':
issue #2 - deny numbers and slashes in first char in variable name
2018-03-30 11:28:50 +00:00
if c == '{' || isValidFirstVarNameRune(rune(c)) {
l.ungetChars(1)
lval.Token(l.newToken(tb[:len(tb)-1]));
split scanner and parser
2018-01-24 16:42:23 +00:00
return T_ENCAPSED_AND_WHITESPACE
issue #2 - deny numbers and slashes in first char in variable name
2018-03-30 11:28:50 +00:00
}
split scanner and parser
2018-01-24 16:42:23 +00:00
case '{':
issue #2 - deny numbers and slashes in first char in variable name
2018-03-30 11:28:50 +00:00
if rune(c) == '$' {
l.ungetChars(1)
lval.Token(l.newToken(tb[:len(tb)-1]));
split scanner and parser
2018-01-24 16:42:23 +00:00
return T_ENCAPSED_AND_WHITESPACE
issue #2 - deny numbers and slashes in first char in variable name
2018-03-30 11:28:50 +00:00
}
split scanner and parser
2018-01-24 16:42:23 +00:00
case '\\':
issue #2 - deny numbers and slashes in first char in variable name
2018-03-30 11:28:50 +00:00
currentChar := l.Last
tb = append(tb, currentChar)
split scanner and parser
2018-01-24 16:42:23 +00:00
c = l.Next();
}
issue #2 - fix template string scanning `l.Prev.Rune` is actually current rune and `c` is next rune.
2018-03-29 13:46:21 +00:00
if rune(c) == '"' {
lval.Token(l.newToken(l.Token()));
return T_ENCAPSED_AND_WHITESPACE
}
issue #2 - deny numbers and slashes in first char in variable name
2018-03-30 11:28:50 +00:00
currentChar = l.Last
tb = append(tb, currentChar)
split scanner and parser
2018-01-24 16:42:23 +00:00
c = l.Next()
issue #2 - fix template string scanning `l.Prev.Rune` is actually current rune and `c` is next rune.
2018-03-29 13:46:21 +00:00
if c == -1 {
break;
}
split scanner and parser
2018-01-24 16:42:23 +00:00
}
<BACKQUOTE>.|[ \t\n\r]
issue #2 - fix backqouted strings now it works as template strings
2018-03-30 15:35:18 +00:00
currentChar := l.Prev
tb := []lex.Char{currentChar}
split scanner and parser
2018-01-24 16:42:23 +00:00
issue #2 - fix backqouted strings now it works as template strings
2018-03-30 15:35:18 +00:00
for {
switch currentChar.Rune {
split scanner and parser
2018-01-24 16:42:23 +00:00
case '$':
issue #2 - fix backqouted strings now it works as template strings
2018-03-30 15:35:18 +00:00
if c == '{' || isValidFirstVarNameRune(rune(c)) {
split scanner and parser
2018-01-24 16:42:23 +00:00
l.ungetChars(1)
issue #2 - fix backqouted strings now it works as template strings
2018-03-30 15:35:18 +00:00
lval.Token(l.newToken(tb[:len(tb)-1]));
split scanner and parser
2018-01-24 16:42:23 +00:00
return T_ENCAPSED_AND_WHITESPACE
issue #2 - fix backqouted strings now it works as template strings
2018-03-30 15:35:18 +00:00
}
split scanner and parser
2018-01-24 16:42:23 +00:00
case '{':
if rune(c) == '$' {
l.ungetChars(1)
lval.Token(l.newToken(tb[:len(tb)-1]));
return T_ENCAPSED_AND_WHITESPACE
issue #2 - fix backqouted strings now it works as template strings
2018-03-30 15:35:18 +00:00
}
split scanner and parser
2018-01-24 16:42:23 +00:00
case '\\':
issue #2 - fix backqouted strings now it works as template strings
2018-03-30 15:35:18 +00:00
currentChar := l.Last
tb = append(tb, currentChar)
split scanner and parser
2018-01-24 16:42:23 +00:00
c = l.Next();
}
issue #2 - fix backqouted strings now it works as template strings
2018-03-30 15:35:18 +00:00
if rune(c) == '`' {
lval.Token(l.newToken(l.Token()));
return T_ENCAPSED_AND_WHITESPACE
}
currentChar = l.Last
tb = append(tb, currentChar)
split scanner and parser
2018-01-24 16:42:23 +00:00
c = l.Next()
issue #2 - fix backqouted strings now it works as template strings
2018-03-30 15:35:18 +00:00
if c == -1 {
break;
}
split scanner and parser
2018-01-24 16:42:23 +00:00
}
<HEREDOC>.|[ \t\n\r]
searchLabel := []byte{}
tb := []lex.Char{}
scanner: fix heredoc
2018-04-05 21:24:00 +00:00
for {
split scanner and parser
2018-01-24 16:42:23 +00:00
if c == -1 {
break;
}
scanner: fix heredoc
2018-04-05 21:24:00 +00:00
nls := 0
split scanner and parser
2018-01-24 16:42:23 +00:00
switch c {
case '\r':
scanner: fix heredoc
2018-04-05 21:24:00 +00:00
nls = 1
c := l.Next()
if c != '\n' {
nls = 0
l.ungetChars(0)
}
fallthrough
case '\n':
split scanner and parser
2018-01-24 16:42:23 +00:00
if bytes.Equal(append(l.charsToBytes(heredocLabel), ';'), searchLabel) {
l.begin(HEREDOC_END)
scanner: fix heredoc
2018-04-05 21:24:00 +00:00
tb = l.ungetChars(len(heredocLabel)+1+nls)
lval.Token(l.newToken(tb));
return T_ENCAPSED_AND_WHITESPACE
split scanner and parser
2018-01-24 16:42:23 +00:00
}
if bytes.Equal(l.charsToBytes(heredocLabel), searchLabel) {
l.begin(HEREDOC_END)
scanner: fix heredoc
2018-04-05 21:24:00 +00:00
tb = l.ungetChars(len(heredocLabel)+nls)
lval.Token(l.newToken(tb));
return T_ENCAPSED_AND_WHITESPACE
split scanner and parser
2018-01-24 16:42:23 +00:00
}
searchLabel = []byte{}
case '$':
c = l.Next();
scanner: fix heredoc
2018-04-05 21:24:00 +00:00
if rune(c) == '{' || isValidFirstVarNameRune(rune(c)) {
split scanner and parser
2018-01-24 16:42:23 +00:00
tb = l.ungetChars(1)
scanner: fix heredoc
2018-04-05 21:24:00 +00:00
lval.Token(l.newToken(tb));
return T_ENCAPSED_AND_WHITESPACE
split scanner and parser
2018-01-24 16:42:23 +00:00
}
l.ungetChars(0)
case '{':
c = l.Next();
if rune(c) == '$' {
tb = l.ungetChars(1)
scanner: fix heredoc
2018-04-05 21:24:00 +00:00
lval.Token(l.newToken(tb));
return T_ENCAPSED_AND_WHITESPACE
split scanner and parser
2018-01-24 16:42:23 +00:00
}
l.ungetChars(0)
case '\\':
c = l.Next();
scanner: fix heredoc
2018-04-05 21:24:00 +00:00
if c == '\n' || c == '\r' {
l.ungetChars(0)
}
split scanner and parser
2018-01-24 16:42:23 +00:00
default:
searchLabel = append(searchLabel, byte(rune(c)))
}
c = l.Next()
}
<STRING_VAR>\${VAR_NAME} lval.Token(l.newToken(l.Token())); return T_VARIABLE
<STRING_VAR>->{VAR_NAME} lval.Token(l.newToken(l.ungetChars(len(l.Token())-2))); return T_OBJECT_OPERATOR
<STRING_VAR>{VAR_NAME} l.popState();lval.Token(l.newToken(l.Token())); return T_STRING
<STRING_VAR>\[ l.pushState(STRING_VAR_INDEX);lval.Token(l.newToken(l.Token())); return Rune2Class(rune(l.TokenBytes(nil)[0]))
<STRING_VAR>.|[ \t\n\r] l.ungetChars(1);l.popState()
<STRING_VAR_INDEX>{LNUM}|{HNUM}|{BNUM} lval.Token(l.newToken(l.Token())); return T_NUM_STRING
<STRING_VAR_INDEX>\${VAR_NAME} lval.Token(l.newToken(l.Token())); return T_VARIABLE
<STRING_VAR_INDEX>{VAR_NAME} lval.Token(l.newToken(l.Token())); return T_STRING
<STRING_VAR_INDEX>\] l.popState(); l.popState();lval.Token(l.newToken(l.Token())); return Rune2Class(rune(l.TokenBytes(nil)[0]))
<STRING_VAR_INDEX>[ \n\r\t\\'#] l.popState(); l.popState();lval.Token(l.newToken(l.Token())); return T_ENCAPSED_AND_WHITESPACE
<STRING_VAR_INDEX>{OPERATORS} lval.Token(l.newToken(l.Token())); return Rune2Class(rune(l.TokenBytes(nil)[0]))
<STRING_VAR_INDEX>. lval.Token(l.newToken(l.Token())); return Rune2Class(rune(l.TokenBytes(nil)[0]))
<STRING_VAR_NAME>{VAR_NAME}[\[\}] l.popState();l.pushState(PHP);lval.Token(l.newToken(l.ungetChars(1))); return T_STRING_VARNAME
<STRING_VAR_NAME>. l.ungetChars(1);l.popState();l.pushState(PHP)
%%
if c, ok := l.Abort(); ok { return int(c) }
goto yyAction
}
Reference in New Issue Copy Permalink