def __init__( self, name=None, **kwds):
name = name or kwds.setdefault( 'context', self.__class__.__name__ )
leng = USINT( context='length' )
leng[None] = cpppo.string_bytes( 'string',
context='string', limit='..length',
initial='.*', decode='iso-8859-1',
terminal=True )
super( SSTRING, self ).__init__( name=name, initial=leng, **kwds )
def __init__(self, name=None, **kwds):
name = name or kwds.setdefault('context', self.__class__.__name__)
leng = USINT(context='length')
leng[None] = cpppo.string_bytes('string',
context='string',
limit='..length',
initial='.*',
decode='iso-8859-1',
terminal=True)
super(SSTRING, self).__init__(name=name, initial=leng, **kwds)
def test_decode():
# Test decode of regexes over bytes data. Operates in raw bytes symbols., works in Python 2/3.
source = cpppo.peekable('π'.encode('utf-8'))
data = cpppo.dotdict()
with cpppo.string_bytes('pi',
initial='.*',
greedy=True,
context='pi',
decode='utf-8') as machine:
for i, (m, s) in enumerate(machine.run(source=source, data=data)):
log.info("%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r",
m.name_centered(), i, s, source.sent, source.peek(), data)
assert i == 3
assert source.sent == 2
assert data.pi == 'π'
if sys.version_info[0] < 3:
# Test regexes over plain string data (no decode required). Force non-unicode (counteracts
# import unicode_literals above). We can't use greenery.lego regexes on unicode data in
# Python 2...
source = cpppo.peekable(str('pi'))
data = cpppo.dotdict()
with cpppo.string('pi', initial='.*', greedy=True,
context='pi') as machine:
for i, (m, s) in enumerate(machine.run(source=source, data=data)):
log.info("%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r",
m.name_centered(), i, s, source.sent, source.peek(),
data)
assert i == 3
assert source.sent == 2
assert data.pi == 'pi'
else:
# Test regexes over Python 3 unicode string data (no decode required). Operates in native
# unicode symbols.
source = cpppo.peekable('π')
data = cpppo.dotdict()
with cpppo.string('pi', initial='.*', greedy=True,
context='pi') as machine:
for i, (m, s) in enumerate(machine.run(source=source, data=data)):
log.info("%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r",
m.name_centered(), i, s, source.sent, source.peek(),
data)
assert i == 2
assert source.sent == 1
assert data.pi == 'π'
source = cpppo.peekable(str('123'))
data = cpppo.dotdict()
with cpppo.integer('value') as machine:
for i, (m, s) in enumerate(machine.run(source=source, data=data)):
log.info("%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r",
m.name_centered(), i, s, source.sent, source.peek(), data)
assert i == 4
assert source.sent == 3
assert data.integer == 123
source = cpppo.peekable('123'.encode('ascii'))
data = cpppo.dotdict()
with cpppo.integer_bytes('value') as machine:
for i, (m, s) in enumerate(machine.run(source=source, data=data)):
log.info("%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r",
m.name_centered(), i, s, source.sent, source.peek(), data)
assert i == 4
assert source.sent == 3
assert data.integer == 123
# Try using a integer (str) parser over bytes data. Works in Python 2, not so much in Python 3
try:
source = cpppo.peekable('123'.encode('ascii'))
data = cpppo.dotdict()
with cpppo.integer('value') as machine:
for i, (m, s) in enumerate(machine.run(source=source, data=data)):
log.info("%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r",
m.name_centered(), i, s, source.sent, source.peek(),
data)
assert i == 4
assert source.sent == 3
assert data.integer == 123
assert sys.version_info[0] < 3, \
"Should have failed in Python3; str/bytes iterator both produce str/int"
except AssertionError:
assert not sys.version_info[0] < 3, \
"Shouldn't have failed in Python2; str/bytes iterator both produce str"
def test_decode():
# Test decode of regexes over bytes data. Operates in raw bytes symbols., works in Python 2/3.
source = cpppo.peekable( 'π'.encode( 'utf-8' ))
data = cpppo.dotdict()
with cpppo.string_bytes( 'pi', initial='.*', greedy=True, context='pi', decode='utf-8' ) as machine:
for i,(m,s) in enumerate( machine.run( source=source, data=data )):
log.info( "%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r", m.name_centered(),
i, s, source.sent, source.peek(), data )
assert i == 3
assert source.sent == 2
assert data.pi == 'π'
if sys.version_info[0] < 3:
# Test regexes over plain string data (no decode required). Force non-unicode (counteracts
# import unicode_literals above). We can't use greenery.lego regexes on unicode data in
# Python 2...
source = cpppo.peekable( str( 'pi' ))
data = cpppo.dotdict()
with cpppo.string( 'pi', initial='.*', greedy=True, context='pi' ) as machine:
for i,(m,s) in enumerate( machine.run( source=source, data=data )):
log.info( "%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r", m.name_centered(),
i, s, source.sent, source.peek(), data )
assert i == 3
assert source.sent == 2
assert data.pi == 'pi'
else:
# Test regexes over Python 3 unicode string data (no decode required). Operates in native
# unicode symbols.
source = cpppo.peekable( 'π' )
data = cpppo.dotdict()
with cpppo.string( 'pi', initial='.*', greedy=True, context='pi' ) as machine:
for i,(m,s) in enumerate( machine.run( source=source, data=data )):
log.info( "%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r", m.name_centered(),
i, s, source.sent, source.peek(), data )
assert i == 2
assert source.sent == 1
assert data.pi == 'π'
source = cpppo.peekable( str( '123' ))
data = cpppo.dotdict()
with cpppo.integer( 'value' ) as machine:
for i,(m,s) in enumerate( machine.run( source=source, data=data )):
log.info( "%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r", m.name_centered(),
i, s, source.sent, source.peek(), data )
assert i == 4
assert source.sent == 3
assert data.integer == 123
source = cpppo.peekable( '123'.encode( 'ascii' ))
data = cpppo.dotdict()
with cpppo.integer_bytes( 'value' ) as machine:
for i,(m,s) in enumerate( machine.run( source=source, data=data )):
log.info( "%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r", m.name_centered(),
i, s, source.sent, source.peek(), data )
assert i == 4
assert source.sent == 3
assert data.integer == 123
# Try using a integer (str) parser over bytes data. Works in Python 2, not so much in Python 3
try:
source = cpppo.peekable( '123'.encode( 'ascii' ))
data = cpppo.dotdict()
with cpppo.integer( 'value' ) as machine:
for i,(m,s) in enumerate( machine.run( source=source, data=data )):
log.info( "%s #%3d -> %10.10s; next byte %3d: %-10.10r: %r", m.name_centered(),
i, s, source.sent, source.peek(), data )
assert i == 4
assert source.sent == 3
assert data.integer == 123
assert sys.version_info[0] < 3, \
"Should have failed in Python3; str/bytes iterator both produce str/int"
except AssertionError:
assert not sys.version_info[0] < 3, \
"Shouldn't have failed in Python2; str/bytes iterator both produce str"
def __init__( self, name=None, **kwds ):
name = name or kwds.setdefault( 'context', self.__class__.__name__ )
# Get the size, and chain remaining machine onto rest. When used as a Route Path, the size
# is padded, so insert a state to drop the pad, and chain rest to that instead.
size = rest = USINT( context='size' )
if self.padsize:
size[True] = rest = octets_drop( 'pad', repeat=1 )
# After capturing each segment__ (pseg), move it onto the path segment list, and loop
pseg = octets_noop( 'type', terminal=True )
# ...segment parsers...
pmov = move_if( 'move', initializer=lambda **kwds: [],
source='..segment__', destination='..segment',
state=pseg )
# Wire each different segment type parser between pseg and pmov
pseg[b'\x28'[0]]= e_8t = octets_drop( 'type', repeat=1 )
e_8t[True] = e_8v = USINT( 'elem_8bit', context='element')
e_8v[None] = pmov
pseg[b'\x29'[0]]= e16t = octets_drop( 'type', repeat=2 )
e16t[True] = e16v = UINT( 'elem16bit', context='element')
e16v[None] = pmov
pseg[b'\x2a'[0]]= e32t = octets_drop( 'type', repeat=2 )
e32t[True] = e32v = UDINT( 'elem32bit', context='element')
e32v[None] = pmov
pseg[b'\x20'[0]]= c_8t = octets_drop( 'type', repeat=1 )
c_8t[True] = c_8v = USINT( 'clas_8bit', context='class')
c_8v[None] = pmov
pseg[b'\x21'[0]]= c16t = octets_drop( 'type', repeat=2 )
c16t[True] = c16v = UINT( 'clas16bit', context='class')
c16v[None] = pmov
pseg[b'\x24'[0]]= i_8t = octets_drop( 'type', repeat=1 )
i_8t[True] = i_8v = USINT( 'inst_8bit', context='instance')
i_8v[None] = pmov
pseg[b'\x25'[0]]= i16t = octets_drop( 'type', repeat=2 )
i16t[True] = i16v = UINT( 'inst16bit', context='instance')
i16v[None] = pmov
pseg[b'\x30'[0]]= a_8t = octets_drop( 'type', repeat=1 )
a_8t[True] = a_8v = USINT( 'attr_8bit', context='attribute')
a_8v[None] = pmov
pseg[b'\x31'[0]]= a16t = octets_drop( 'type', repeat=2 )
a16t[True] = a16v = UINT( 'attr16bit', context='attribute')
a16v[None] = pmov
pseg[b'\x91'[0]]= symt = octets_drop( 'type', repeat=1 )
symt[True] = syml = USINT( 'sym_len', context='symbolic.length' )
syml[None] = symv = cpppo.string_bytes(
'symbolic', context='symbolic', limit='.length',
initial='.*', decode='iso-8859-1' )
# An odd-length ANSI Extended Symbolic name means an odd total. Pad
symo = octets_drop( 'pad', repeat=1 )
symo[None] = pmov
symv[None] = cpppo.decide( 'odd',
predicate=lambda path=None, data=None, **kwds: len( data[path].symbolic ) % 2,
state=symo )
symv[None] = pmov
# Route Path port/link-address. See Vol 1-3.13, Table C-1.3 Port Segment Encoding.
# segment: 0b000spppp
# |\\\\+-> port number 0x01-0x0E; 0x0F=>extended
# |
# +------> link size+address; 0=>numeric, 1=>size+string
#
def port_fix( path=None, data=None, **kwds ):
"""Discard port values about 0x0F; return True (transition) if remaining port value is 0x0F
(Optional Extended port)"""
data[path].port &= 0x0F
if data[path].port == 0x0F:
# Port is extended; discard and prepare to collect new port number
data[path].port = cpppo.dotdict()
return True
# Port is OK; don't transition
return False
# [01-0E][LL] port 01-0E, link-address #LL
pseg[b'\x01'[0]]= pnum = USINT( 'port_num', context='port' )
pseg[b'\x02'[0]] = pnum
pseg[b'\x03'[0]] = pnum
pseg[b'\x04'[0]] = pnum
pseg[b'\x05'[0]] = pnum
pseg[b'\x06'[0]] = pnum
pseg[b'\x07'[0]] = pnum
pseg[b'\x08'[0]] = pnum
pseg[b'\x09'[0]] = pnum
pseg[b'\x0a'[0]] = pnum
pseg[b'\x0b'[0]] = pnum
pseg[b'\x0c'[0]] = pnum
pseg[b'\x0d'[0]] = pnum
pseg[b'\x0e'[0]] = pnum
# [0F][PPPP][LL] port 0xPPPP, link-address 0xLL
pseg[b'\x0f'[0]] = pnum
# A big port#; re-scan a UINT into .port (won't work 'til port_fix is called)
pnbg = UINT( 'port_nbg', context='port' )
pnbg[True] = pnlk = USINT( 'link_num', context='link' )
# Fix the port#; if 0x0F, setup for extended port and transition to pnbg. Otherwise,
# (not extended port), just go the the port numeric link.
pnum[None] = cpppo.decide( 'port_nfix', predicate=port_fix,
state=pnbg )
pnum[None] = pnlk
pnlk[None] = pmov # and done; move segment, get next
# [11-1E][SS]'123.123.123.123'[00] port 0x01-0E, link address '123.123.123.123' (pad if size 0xSS odd)
pseg[b'\x11'] = padr = USINT( 'port_adr', context='port' )
pseg[b'\x12'[0]] = padr
pseg[b'\x13'[0]] = padr
pseg[b'\x14'[0]] = padr
pseg[b'\x15'[0]] = padr
pseg[b'\x16'[0]] = padr
pseg[b'\x17'[0]] = padr
pseg[b'\x18'[0]] = padr
pseg[b'\x19'[0]] = padr
pseg[b'\x1a'[0]] = padr
pseg[b'\x1b'[0]] = padr
pseg[b'\x1c'[0]] = padr
pseg[b'\x1d'[0]] = padr
pseg[b'\x1e'[0]] = padr
# [1F][SS][PPPP]'123.123.123.123'[00] port 0xPPPP, link address '123.123.123.123' (pad if size SS odd)
pseg[b'\x1f'[0]] = padr
# Harvest the addresses into .link
adrv = cpppo.string_bytes(
'link_add', context='link', limit='.length',
initial='.*', decode='iso-8859-1' )
# An odd-length link address means an odd total. Pad
adro = octets_drop( 'link_pad', repeat=1 )
adro[None] = pmov
adrv[None] = cpppo.decide( 'link_odd',
predicate=lambda path=None, data=None, **kwds: len( data[path+'.link'] ) % 2,
state=adro )
adrv[None] = pmov
# A big port#; re-scan a UINT into .port (won't work 'til port_fix is called)
pabg = UINT( 'port_abg', context='port' )
pabg[None] = adrv
#
padr[True] = adrl = USINT( 'link_len', context='link.length' )
adrl[None] = cpppo.decide( 'port_afix', predicate=port_fix,
state=pabg )
adrl[None] = adrv
# Parse all segments in a sub-dfa limited by the parsed path.size (in words; double)
rest[None] = cpppo.dfa( 'each', context='segment__',
initial=pseg, terminal=True,
limit=lambda path=None, data=None, **kwds: data[path+'..size'] * 2 )
super( EPATH, self ).__init__( name=name, initial=size, **kwds )
def __init__(self, name=None, **kwds):
name = name or kwds.setdefault('context', self.__class__.__name__)
# Get the size, and chain remaining machine onto rest. When used as a Route Path, the size
# is padded, so insert a state to drop the pad, and chain rest to that instead.
size = rest = USINT(context='size')
if self.padsize:
size[True] = rest = octets_drop('pad', repeat=1)
# After capturing each segment__ (pseg), move it onto the path segment list, and loop
pseg = octets_noop('type', terminal=True)
# ...segment parsers...
pmov = move_if('move',
initializer=lambda **kwds: [],
source='..segment__',
destination='..segment',
state=pseg)
# Wire each different segment type parser between pseg and pmov
pseg[b'\x28'[0]] = e_8t = octets_drop('type', repeat=1)
e_8t[True] = e_8v = USINT('elem_8bit', context='element')
e_8v[None] = pmov
pseg[b'\x29'[0]] = e16t = octets_drop('type', repeat=2)
e16t[True] = e16v = UINT('elem16bit', context='element')
e16v[None] = pmov
pseg[b'\x2a'[0]] = e32t = octets_drop('type', repeat=2)
e32t[True] = e32v = UDINT('elem32bit', context='element')
e32v[None] = pmov
pseg[b'\x20'[0]] = c_8t = octets_drop('type', repeat=1)
c_8t[True] = c_8v = USINT('clas_8bit', context='class')
c_8v[None] = pmov
pseg[b'\x21'[0]] = c16t = octets_drop('type', repeat=2)
c16t[True] = c16v = UINT('clas16bit', context='class')
c16v[None] = pmov
pseg[b'\x24'[0]] = i_8t = octets_drop('type', repeat=1)
i_8t[True] = i_8v = USINT('inst_8bit', context='instance')
i_8v[None] = pmov
pseg[b'\x25'[0]] = i16t = octets_drop('type', repeat=2)
i16t[True] = i16v = UINT('inst16bit', context='instance')
i16v[None] = pmov
pseg[b'\x30'[0]] = a_8t = octets_drop('type', repeat=1)
a_8t[True] = a_8v = USINT('attr_8bit', context='attribute')
a_8v[None] = pmov
pseg[b'\x31'[0]] = a16t = octets_drop('type', repeat=2)
a16t[True] = a16v = UINT('attr16bit', context='attribute')
a16v[None] = pmov
pseg[b'\x91'[0]] = symt = octets_drop('type', repeat=1)
symt[True] = syml = USINT('sym_len', context='symbolic.length')
syml[None] = symv = cpppo.string_bytes('symbolic',
context='symbolic',
limit='.length',
initial='.*',
decode='iso-8859-1')
# An odd-length ANSI Extended Symbolic name means an odd total. Pad
symo = octets_drop('pad', repeat=1)
symo[None] = pmov
symv[None] = cpppo.decide('odd',
predicate=lambda path=None, data=None, **
kwds: len(data[path].symbolic) % 2,
state=symo)
symv[None] = pmov
# Route Path port/link-address. See Vol 1-3.13, Table C-1.3 Port Segment Encoding.
# segment: 0b000spppp
# |\\\\+-> port number 0x01-0x0E; 0x0F=>extended
# |
# +------> link size+address; 0=>numeric, 1=>size+string
#
def port_fix(path=None, data=None, **kwds):
"""Discard port values about 0x0F; return True (transition) if remaining port value is 0x0F
(Optional Extended port)"""
data[path].port &= 0x0F
if data[path].port == 0x0F:
# Port is extended; discard and prepare to collect new port number
data[path].port = cpppo.dotdict()
return True
# Port is OK; don't transition
return False
# [01-0E][LL] port 01-0E, link-address #LL
pseg[b'\x01'[0]] = pnum = USINT('port_num', context='port')
pseg[b'\x02'[0]] = pnum
pseg[b'\x03'[0]] = pnum
pseg[b'\x04'[0]] = pnum
pseg[b'\x05'[0]] = pnum
pseg[b'\x06'[0]] = pnum
pseg[b'\x07'[0]] = pnum
pseg[b'\x08'[0]] = pnum
pseg[b'\x09'[0]] = pnum
pseg[b'\x0a'[0]] = pnum
pseg[b'\x0b'[0]] = pnum
pseg[b'\x0c'[0]] = pnum
pseg[b'\x0d'[0]] = pnum
pseg[b'\x0e'[0]] = pnum
# [0F][PPPP][LL] port 0xPPPP, link-address 0xLL
pseg[b'\x0f'[0]] = pnum
# A big port#; re-scan a UINT into .port (won't work 'til port_fix is called)
pnbg = UINT('port_nbg', context='port')
pnbg[True] = pnlk = USINT('link_num', context='link')
# Fix the port#; if 0x0F, setup for extended port and transition to pnbg. Otherwise,
# (not extended port), just go the the port numeric link.
pnum[None] = cpppo.decide('port_nfix', predicate=port_fix, state=pnbg)
pnum[None] = pnlk
pnlk[None] = pmov # and done; move segment, get next
# [11-1E][SS]'123.123.123.123'[00] port 0x01-0E, link address '123.123.123.123' (pad if size 0xSS odd)
pseg[b'\x11'] = padr = USINT('port_adr', context='port')
pseg[b'\x12'[0]] = padr
pseg[b'\x13'[0]] = padr
pseg[b'\x14'[0]] = padr
pseg[b'\x15'[0]] = padr
pseg[b'\x16'[0]] = padr
pseg[b'\x17'[0]] = padr
pseg[b'\x18'[0]] = padr
pseg[b'\x19'[0]] = padr
pseg[b'\x1a'[0]] = padr
pseg[b'\x1b'[0]] = padr
pseg[b'\x1c'[0]] = padr
pseg[b'\x1d'[0]] = padr
pseg[b'\x1e'[0]] = padr
# [1F][SS][PPPP]'123.123.123.123'[00] port 0xPPPP, link address '123.123.123.123' (pad if size SS odd)
pseg[b'\x1f'[0]] = padr
# Harvest the addresses into .link
adrv = cpppo.string_bytes('link_add',
context='link',
limit='.length',
initial='.*',
decode='iso-8859-1')
# An odd-length link address means an odd total. Pad
adro = octets_drop('link_pad', repeat=1)
adro[None] = pmov
adrv[None] = cpppo.decide('link_odd',
predicate=lambda path=None, data=None, **
kwds: len(data[path + '.link']) % 2,
state=adro)
adrv[None] = pmov
# A big port#; re-scan a UINT into .port (won't work 'til port_fix is called)
pabg = UINT('port_abg', context='port')
pabg[None] = adrv
#
padr[True] = adrl = USINT('link_len', context='link.length')
adrl[None] = cpppo.decide('port_afix', predicate=port_fix, state=pabg)
adrl[None] = adrv
# Parse all segments in a sub-dfa limited by the parsed path.size (in words; double)
rest[None] = cpppo.dfa('each',
context='segment__',
initial=pseg,
terminal=True,
limit=lambda path=None, data=None, **kwds: data[
path + '..size'] * 2)
super(EPATH, self).__init__(name=name, initial=size, **kwds)
