- Much improved parser.

This commit is contained in:
ccremers 2005-11-29 14:16:36 +00:00
parent 6543a8f659
commit dbc5d62ef6
3 changed files with 131 additions and 249 deletions

View File

@ -4,7 +4,6 @@ set cpo&vim
imap <C-F3> :call Cream_replacemulti()
map! <S-Insert> <MiddleMouse>
nnoremap <silent>  :exe 'silent! normal! za'.(foldlevel('.')?'':'l')
map "," :!./parser.py
map , :make
nmap <C-F3> :call Cream_replacemulti()
vmap <C-F3> :call Cream_replacemulti()
@ -31,7 +30,7 @@ set incsearch
set nojoinspaces
set laststatus=2
set listchars=tab:>.,trail:$
set makeprg=./if2spdl.py
set makeprg=./parser.py
set printencoding=ascii
set ruler
set scrolloff=2
@ -43,7 +42,6 @@ set viminfo=%,'50,\"1000,:100,f1,n~/.viminfo
set visualbell
set whichwrap=<,>,h,l
set wildmenu
set window=30
let s:so_save = &so | let s:siso_save = &siso | set so=0 siso=0
let v:this_session=expand("<sfile>:p")
silent only
@ -52,25 +50,17 @@ if expand('%') == '' && !&modified && line('$') <= 1 && getline(1) == ''
let s:wipebuf = bufnr('%')
endif
set shortmess=aoO
badd +9 parser.py
badd +0 NSPK_LOWE.if1
badd +8 NSPK_LOWE.if
badd +6 if2spdl.py
badd +0 generator.py
args parser.py
edit generator.py
badd +10 if2spdl.py
badd +0 parser.py
args if2spdl.py
edit parser.py
set splitbelow splitright
wincmd _ | wincmd |
split
1wincmd k
wincmd w
set nosplitbelow
set nosplitright
wincmd t
set winheight=1 winwidth=1
exe '1resize ' . ((&lines * 15 + 15) / 31)
exe '2resize ' . ((&lines * 13 + 15) / 31)
argglobal
setlocal noarabic
setlocal autoindent
setlocal autoread
setlocal nobinary
@ -116,6 +106,7 @@ setlocal indentexpr=GetPythonIndent(v:lnum)
setlocal indentkeys=0{,0},:,0#,!^F,o,O,e,<:>,=elif,=except
setlocal noinfercase
setlocal iskeyword=@,48-57,_,192-255
setlocal keymap=
setlocal keywordprg=
setlocal nolinebreak
setlocal nolisp
@ -133,6 +124,8 @@ setlocal nopreserveindent
setlocal nopreviewwindow
setlocal quoteescape=\\
setlocal noreadonly
setlocal norightleft
setlocal rightleftcmd=search
setlocal noscrollbind
setlocal shiftwidth=8
setlocal noshortname
@ -157,115 +150,16 @@ setlocal nowinfixheight
setlocal wrap
setlocal wrapmargin=0
silent! normal! zE
let s:l = 4 - ((3 * winheight(0) + 7) / 15)
let s:l = 246 - ((33 * winheight(0) + 27) / 55)
if s:l < 1 | let s:l = 1 | endif
exe s:l
normal! zt
4
246
normal! 01l
wincmd w
argglobal
edit if2spdl.py
setlocal autoindent
setlocal autoread
setlocal nobinary
setlocal bufhidden=
setlocal buflisted
setlocal buftype=
setlocal nocindent
setlocal cinkeys=0{,0},0),:,0#,!^F,o,O,e
setlocal cinoptions=
setlocal cinwords=if,else,while,do,for,switch
setlocal comments=s1:/*,mb:*,ex:*/,://,b:#,:%,:XCOMM,n:>,fb:-
setlocal commentstring=/*%s*/
setlocal complete=.,w,b,u,t,i
setlocal completefunc=
setlocal nocopyindent
setlocal define=
setlocal dictionary=
setlocal nodiff
setlocal equalprg=
setlocal errorformat=
setlocal noexpandtab
if &filetype != 'python'
setlocal filetype=python
endif
setlocal foldcolumn=0
setlocal foldenable
setlocal foldexpr=0
setlocal foldignore=#
setlocal foldlevel=0
setlocal foldmarker={{{,}}}
setlocal foldmethod=manual
setlocal foldminlines=1
setlocal foldnestmax=20
setlocal foldtext=foldtext()
setlocal formatoptions=cqrt
setlocal formatlistpat=^\\s*\\d\\+[\\]:.)}\\t\ ]\\s*
setlocal grepprg=
setlocal iminsert=2
setlocal imsearch=2
setlocal include=
setlocal includeexpr=
setlocal indentexpr=GetPythonIndent(v:lnum)
setlocal indentkeys=0{,0},:,0#,!^F,o,O,e,<:>,=elif,=except
setlocal noinfercase
setlocal iskeyword=@,48-57,_,192-255
setlocal keywordprg=
setlocal nolinebreak
setlocal nolisp
setlocal nolist
setlocal makeprg=
setlocal matchpairs=(:),{:},[:]
setlocal modeline
setlocal modifiable
setlocal nrformats=octal,hex
setlocal nonumber
setlocal numberwidth=4
setlocal omnifunc=
setlocal path=
setlocal nopreserveindent
setlocal nopreviewwindow
setlocal quoteescape=\\
setlocal noreadonly
setlocal noscrollbind
setlocal shiftwidth=8
setlocal noshortname
setlocal nosmartindent
setlocal softtabstop=0
setlocal nospell
setlocal spellcapcheck=[.?!]\\_[\\])'\"\ \ ]\\+
setlocal spellfile=
setlocal spelllang=en
setlocal statusline=
setlocal suffixesadd=
setlocal swapfile
setlocal synmaxcol=3000
if &syntax != 'python'
setlocal syntax=python
endif
setlocal tabstop=8
setlocal tags=
setlocal textwidth=72
setlocal thesaurus=
setlocal nowinfixheight
setlocal wrap
setlocal wrapmargin=0
silent! normal! zE
let s:l = 13 - ((9 * winheight(0) + 6) / 13)
if s:l < 1 | let s:l = 1 | endif
exe s:l
normal! zt
13
normal! 0
wincmd w
2wincmd w
if exists('s:wipebuf')
exe 'bwipe ' . s:wipebuf
endif
unlet! s:wipebuf
exe '1resize ' . ((&lines * 15 + 15) / 31)
exe '2resize ' . ((&lines * 13 + 15) / 31)
set winheight=1 winwidth=20 shortmess=at
let s:sx = expand("<sfile>:p:r")."x.vim"
if file_readable(s:sx)

14
scripts/if2spdl/notes.txt Normal file
View File

@ -0,0 +1,14 @@
Notes
-----
Regarding the AVISPA IF report:
- Messages do not seem to contain Varable' as an option, seems to be
flaw in the BNF.
- State has no BNF definition in the second half, it seems.
- Is Fact well-defined?
- Authenticate has no parameters in it (only constants)
STSecrecy
matching_request

View File

@ -4,15 +4,23 @@
# http://pyparsing.sourceforge.net/
from pyparsing import Literal, alphas, nums, Word, oneOf, Or, Group, \
restOfLine, Forward, Optional, delimitedList
restOfLine, Forward, Optional, delimitedList, alphanums,\
OneOrMore
import Term
typedversion = False
# Generate atom parser
# Markers:
#
# TODO stuff that still needs to be done.
# DEVIANT stuff that deviates from the original BNF specs in the
# paper.
# TEST test things, remove later
# Generate parser
#
# Takes a list of tokens, returns
def atomsParser ():
def ruleParser ():
global typedversion
# ------------------------------------------------------
@ -37,30 +45,41 @@ def atomsParser ():
# Typeinfo/Constant
TypeInfo = oneOf ("mr nonce pk sk fu table")
TypeInfo.setParseAction(lambda s,l,t: [ "typeinfo", Term.TermConstant(t[0]) ])
Const = Word(alphas,Alfabet)
Const.setParseAction(lambda s,l,t: [ "constant", Term.TermConstant(t[0]) ])
Constant = Word(alphas,Alfabet)
Constant.setParseAction(lambda s,l,t: [ "constant", Term.TermConstant(t[0]) ])
# Time
nTime = Group(Number)
nTime.setParseAction(lambda s,l,t: ["n", t[0] ])
xTime = Literal("xTime")
xTime.setParseAction(lambda s,l,t: ["x", 0 ])
xTime.setParseAction(lambda s,l,t: ["x", t[0] ])
sTime = Literal("s").suppress() + lbr + Group(Number) + rbr
sTime.setParseAction(lambda s,l,t: ["s", t[0] ])
Time = Or([nTime,xTime,sTime])
Time.setParseAction(lambda s,l,t: ["time", t[0],t[1] ])
# Const
Const = Forward()
Const << Or ([ Constant, Literal("c") + lbr + Constant + comma + Time + rbr, Literal("c(ni,ni)") ])
# Two versions
Variable = Word("x",Alfabet)
Variable.setParseAction(lambda s,l,t: [ "v", Term.TermVariable(t[0],None) ])
if typedversion:
Variable = TypeInfo + "(" + Variable + ")"
Variable = TypeInfo + lbr + Variable + rbr
# Optional prime
optprime = Optional(Literal("'"))
# Atomic
Atomic = Or([ TypeInfo + lbr + Const + rbr, Variable])
## DEVIANT : below there is an optprime after the atom. This
## is not in the BNF.
Atomic = Or([ TypeInfo + lbr + Const + rbr, Variable]) + optprime
### TEST
print Time.parseString("s(25)")
#print Time.parseString("s(25)")
#print Variable.parseString("xCas")
#print Atomic.parseString("nonce(Koen)")
# ------------------------------------------------------
# Messages
@ -69,8 +88,6 @@ def atomsParser ():
# Base forward declaration
Message = Forward()
# Optional prime
optprime = Optional(Literal("'"))
# Agents etc
Agent = Or ([Literal("mr") + lbr + Const + rbr, Variable])
@ -91,7 +108,10 @@ def atomsParser ():
Concatenation = Literal("c") + lbr + Message + comma + Message + rbr
Composed = Or([ Concatenation, SymmetricCypher, XOR,
PublicCypher, Function, KeyTable, KeyTableApp ])
Message = Or ([Composed, Atomic])
Message << Or ([Composed, Atomic])
### TEST
#print Message.parseString("nonce(c(Na,xTime))")
# ------------------------------------------------------
# Model of honest agents
@ -99,18 +119,20 @@ def atomsParser ():
Boolean = Or ([ Literal("true"), Literal("false"), Variable ])
Session = Forward()
Session = Or ([ Literal("s") + lbr + Session + rbr, Number, Variable ])
MsgList = Forward()
Session << Or ([ Literal("s") + lbr + Session + rbr, Number, Variable ])
MsgEtc = Literal("etc")
MsgList = Forward()
MsgComp = Literal("c") + lbr + Message + comma + MsgList + rbr
MsgList = Or ([ MsgEtc, Variable, MsgComp ])
MsgList << Or ([ MsgEtc, Variable, MsgComp ])
Step = Or ([ Number, Variable ])
### TEST
print Message.parseString("xKb")
print MsgList.parseString("etc")
print MsgComp.parseString("c(xKb,etc)")
print MsgList.parseString("c(xA,c(xB,c(xKa,c(xKa',c(xKb,etc)))))")
#print Message.parseString("xKb")
#print MsgList.parseString("etc")
#print MsgList.parseString("c(xKb,etc)")
#print MsgList.parseString("c(xA,c(xB,c(xKa,c(xKa',c(xKb,etc)))))")
# Principal fact
Principal = Literal("w") + lbr + Step + comma + Agent + comma + Agent + comma + MsgList + comma + MsgList + comma + Boolean + comma + Session + rbr
@ -119,148 +141,100 @@ def atomsParser ():
MessageFact = Literal("m") + lbr + Step + comma + Agent + comma + Agent + comma + Agent + comma + Message + comma + Session + rbr
# Goal fact
GoalFact = Literal ("nogniet")
GoalState = Literal ("nogniet")
Correspondence = Principal + dot + Principal
Secret = Literal("secret") + lbr + Message + Literal("f") + lbr + Session + rbr + rbr
Secrecy = Literal("secret") + lbr + Literal("xsecret") + comma + Literal("f") + lbr + Session + rbr + rbr + dot + Literal("i") + lbr + Literal("xsecret") + rbr
Give = Literal("give") + lbr + Message + Literal("f") + lbr + Session + rbr + rbr
STSecrecy = Literal("give(xsecret,f(xc)).secret(xsecret,f(xc))") + implies + Literal("i(xsecret)")
Witness = Literal("witness") + lbr + Agent + comma + Agent + comma + Constant + comma + Message + rbr
Request = Literal("request") + lbr + Agent + comma + Agent + comma + Constant + comma + Message + rbr
Authenticate = Literal("request") + lbr + Agent + comma + Agent + comma + Constant + comma + Message + rbr
GoalFact = Or ([ Correspondence, Secrecy, STSecrecy, Authenticate ])
GoalState = Or ([ Secret, Give, Witness, Request ])
# Facts and states
Fact = Or ([ Principal, MessageFact ])
State = Group(delimitedList (Fact, "."))
Fact = Or ([ Principal, MessageFact, GoalFact ]) ## Not well defined in BNF
State = Group(delimitedList (Fact, ".")) ## From initial part of document, not in detailed BNF
# Rules
mr1 = Literal("h") + lbr + Literal("s") + lbr + Literal("xTime") + rbr + rbr + dot + State
MFPrincipal = Or ([ MessageFact + dot + Principal, Principal ])
mr1 = Literal("h") + lbr + Literal("s") + lbr + Literal("xTime") + rbr + rbr + dot + MFPrincipal
mr2 = implies
mr3 = Literal("h") + lbr + Literal("xTime") + rbr + dot + MessageFact + dot + Principal + dot + GoalFact + eol
MessageRule = mr1 + eol + mr2 + eol + mr3 + eol
InitialState = Literal("h") + lbr + Literal("xTime") + rbr + dot + State + eol
mr3 = Literal("h") + lbr + Literal("xTime") + rbr + dot + MFPrincipal + Optional(dot + delimitedList(GoalFact, "."))
MessageRule = mr1 + mr2 + mr3 ## DEVIANT : BNF requires newlines
InitialState = Literal("h") + lbr + Literal("xTime") + rbr + dot + State ## DEVIANT : BNF requires newlines
# Intruder
IntruderRule = Literal("nogniet")
# Simplification
SimplificationRule = Literal("nogniet")
f_simplif = Literal("f") + lbr + Literal("s") + lbr + Literal ("xc") + rbr + rbr + implies + Literal("f") + lbr + Literal("xc") + rbr ## DEVIANT : EOL removed
matching_request = Witness + dot + Request + implies
no_auth_intruder = Request + implies
SimplificationRule = Or ([ f_simplif, matching_request, no_auth_intruder ])
# Compose all rules
Rule = Or([ InitialState, MessageRule, IntruderRule, GoalState, SimplificationRule ])
return Rule
print Rule.parseFile("test.if")
# IFParser
# Does not work for the first line (typed/untyped)
# Depends on ruleParser
def ifParser():
def ifParse (str):
# Tokens
lbr = Literal("(").suppress()
rbr = Literal(")").suppress()
comma = Literal(",").suppress()
hash = Literal("#").suppress()
equ = Literal("=").suppress()
implies = Literal("=>").suppress()
# Functions to construct tuples etc
def bracket(x):
return lbr + x + rbr
def ntup(n):
x = Message
while n > 1:
x = x + comma + Message
n = n - 1
return x
def btup(n):
return bracket(ntup(n))
def funcy(x,y):
return x + bracket(y)
def ftup(x,n):
return funcy(x, ntup(n))
# Message section
Alfabet= alphas+nums+"_$"
Variable = Word("x",Alfabet).setParseAction(lambda s,l,t: [ Term.TermVariable(t[0],None) ])
Constant = Word(alphas,Alfabet).setParseAction(lambda s,l,t: [ Term.TermConstant(t[0]) ])
Number = Word(nums).setParseAction(lambda s,l,t: [ Term.TermConstant(t[0]) ])
Basic = MatchFirst([ Variable, Constant, Number ])
# Message definition is recursive
Message = Forward()
def parseType(s,l,t):
if t[0][0] == "pk":
# Public key thing, that's not really a type for
# us but a function
return [Term.TermEncrypt(t[0][1], t[0][0]) ]
term = t[0][1]
term.setType(t[0][0])
return [term]
TypeInfo = oneOf ("mr nonce pk sk fu table").setParseAction(lambda s,l,t: [ Term.TermConstant(t[0]) ])
TypeMsg = Group(TypeInfo + lbr + Message + rbr).setParseAction(parseType)
def parseCrypt(s,l,t):
# Crypto types are ignored for now
type = t[0][0]
if type == "c":
return [Term.TermTuple( t[0][1],t[0][2] ) ]
return [Term.TermEncrypt(t[0][2],t[0][1])]
CryptOp = oneOf ("crypt scrypt c funct rcrypt tb")
CryptMsg = Group(CryptOp + lbr + Message + comma + Message + rbr).setParseAction(parseCrypt)
def parseSMsg(s,l,t):
return [Term.TermEncrypt(t[0][1],Term.Termconstant("succ") )]
SMsg = Group(Literal("s") + lbr + Message + rbr)
def parsePrime(s,l,t):
# for now, we simply ignore the prime (')
return [t[0][0]]
Message << Group(Or ([TypeMsg, CryptMsg, SMsg, Basic]) + Optional(Literal("'"))).setParseAction(parsePrime)
# Fact section
Request = Group("request" + btup(4))
Witness = Group("witness" + btup(4))
Give = Group("give" + lbr + Message + comma + ftup(Literal("f"),
1) + rbr)
Secret = Group("secret" + lbr + Message + comma +
ftup(Literal("f"),1) + rbr)
TimeFact = Group(ftup (Literal("h"), 1))
IntruderKnowledge = Group(ftup (Literal("i"), 1))
MessageFact = Group(ftup(Literal("m"),6))
Principal = Group(ftup(Literal("w"), 7))
Fact = Principal | MessageFact | IntruderKnowledge | TimeFact | Secret | Give | Witness | Request
#State = Fact + OptioZeroOrMore ("." + Fact)
State = Group(delimitedList (Fact, "."))
equal = Literal("=").suppress()
# Rules and labels
rulename = Word (alphanums + "_")
rulecategory = oneOf("Protocol_Rules Invariant_Rules Decomposition_Rules Intruder_Rules Init Goal")
label = hash + "lb" + equ + rulename + comma + "type" + equ + rulecategory
rule = Group(State + Optional(implies + State))
labeledrule = Group(label + rule)
typeflag = hash + "option" + equ + oneOf ("untyped","typed")
label = hash + Literal("lb") + equal + rulename + comma + Literal("type") + equal + rulecategory
labeledrule = Group(label + ruleParser())
# A complete file
iffile = typeflag + Group(OneOrMore(labeledrule))
parser = iffile
parser = Group(OneOrMore(labeledrule))
parser.ignore("##" + restOfLine)
return parser.parseString(str)
return parser
# Determine (un)typedness from this line
def typeSwitch(line):
try:
global typedversion
typeflag = Literal("#") + "option" + Literal("=") + oneOf ("untyped","typed")
res = typeflag.parseString(line)
if res[3] == "untyped":
typedversion = False
elif res[3] == "typed":
typeversion = True
else:
print "Cannot determine whether typed or untyped."
raise ParseException
except:
print "Unexpected error while determining (un)typedness of the line", line
# Parse an entire file, including the first one
def linesParse(lines):
typeSwitch(lines[0])
parser = ifParser()
result = parser.parseString("".join( lines[1:]))
for x in result:
print x
# Main code
def main():
global typedversion
file = open("NSPK_LOWE.if", "r")
linesParse(file.readlines())
file.close()
typedversion = False
atomsParser()
if __name__ == '__main__':
main()