def test_parser_edge_cases():
# Our parser implementation is general enough that
# some of its branches are not being exercised by our regular tests,
# so I had to come up with these contrived examples to test them.
p = many(rfc7230.tchar) > named(u'p')
p1 = '1' * p > named(u'p1')
p2 = '11' * p * skip('\n') > named(u'p2')
assert parse(p1 | p2, b'11abc') == (u'1', [u'1', u'a', u'b', u'c'])
assert parse(p1 | p2, b'11abc\n') == (u'11', [u'a', u'b', u'c'])
p = recursive() > named(u'p')
p.rec = (rfc7230.tchar * p | subst(None) << empty)
assert parse(p, b'abc') == (u'a', (u'b', (u'c', None)))
p = literal('ab') > named(u'p')
p0 = subst(u'') << empty | p > named(u'p0')
p1 = 'xab' * p0 > named(u'p1')
p2 = 'x' * string(p0) * '!' > named(u'p2')
assert parse(p1 | p2, b'xabab') == (u'xab', u'ab')
assert parse(p1 | p2, b'xabab!') == (u'x', u'abab', u'!')
p = empty | literal('a') > named(u'p')
p0 = p * 'x' > named(u'x')
assert parse(p0, b'x') == u'x'
def test_parser_edge_cases():
# Our parser implementation is general enough that
# some of its branches are not being exercised by our regular tests,
# so I had to come up with these contrived examples to test them.
p = many(rfc7230.tchar) > named(u'p')
p1 = '1' * p > named(u'p1')
p2 = '11' * p * skip('\n') > named(u'p2')
assert parse(p1 | p2, b'11abc') == (u'1', [u'1', u'a', u'b', u'c'])
assert parse(p1 | p2, b'11abc\n') == (u'11', [u'a', u'b', u'c'])
p = recursive() > named(u'p')
p.rec = (rfc7230.tchar * p | subst(None) << empty)
assert parse(p, b'abc') == (u'a', (u'b', (u'c', None)))
p = literal('ab') > named(u'p')
p0 = subst(u'') << empty | p > named(u'p0')
p1 = 'xab' * p0 > named(u'p1')
p2 = 'x' * string(p0) * '!' > named(u'p2')
assert parse(p1 | p2, b'xabab') == (u'xab', u'ab')
assert parse(p1 | p2, b'xabab!') == (u'x', u'abab', u'!')
p = empty | literal('a') > named(u'p')
p0 = p * 'x' > named(u'x')
assert parse(p0, b'x') == u'x'
def media_range(no_q=False):
return Parametrized << (
(
literal('*/*') |
type_ + '/' + '*' |
MediaType << type_ + '/' + subtype
) *
(
MultiDict << many(
skip(OWS * ';' * OWS) *
parameter(exclude=['q'] if no_q else [])
)
)
) > named(u'media-range', RFC(7231), is_pivot=True)
# -*- coding: utf-8; -*-
from httpolice.citation import RFC
from httpolice.parse import auto, fill_names, many, maybe, pivot, skip, string1
from httpolice.structure import HSTSDirective, Parametrized
from httpolice.syntax.common import DIGIT
from httpolice.syntax.rfc7230 import OWS, quoted_string, token
# This has been slightly adapted to the rules of RFC 7230.
# The ``OWS`` are derived from the "implied ``*LWS``" requirement.
directive_name = HSTSDirective << token > auto
directive_value = token | quoted_string > auto
directive = Parametrized << (
directive_name * maybe(skip(OWS * '=' * OWS) * directive_value)) > pivot
def _collect_elements(xs):
return [elem for elem in xs if elem is not None]
Strict_Transport_Security = _collect_elements << (
maybe(directive) % many(skip(OWS * ';' * OWS) * maybe(directive))) > pivot
max_age_value = int << string1(DIGIT) > pivot
fill_names(globals(), RFC(6797))
correct_encoded_id += pct_encode(c, safe='').upper()
if encoded_id != correct_encoded_id:
complain(1256, actual=encoded_id, correct=correct_encoded_id)
return decoded_id
protocol_id = _check_protocol_id << token > pivot
@can_complain
def _check_alt_authority(complain, value):
return parse(value, maybe_str(uri_host) + ':' + port, complain, 1257,
authority=value)
alt_authority = _check_alt_authority << quoted_string > pivot
alternative = protocol_id * skip('=') * alt_authority > pivot
parameter = ((AltSvcParam << token) *
skip('=') * (token | quoted_string)) > pivot
alt_value = Parametrized << (
alternative *
(MultiDict << many(skip(OWS * ';' * OWS) * parameter))) > pivot
Alt_Svc = clear | comma_list1(alt_value) > pivot
ma = delta_seconds > pivot
persist = subst(True) << literal('1') > pivot
Alt_Used = uri_host + maybe_str(':' + port) > pivot
fill_names(globals(), RFC(7838))
protocol_id = _check_protocol_id << token > pivot
@can_complain
def _check_alt_authority(complain, value):
return parse(value,
maybe_str(uri_host) + ':' + port,
complain,
1257,
authority=value)
alt_authority = _check_alt_authority << quoted_string > pivot
alternative = protocol_id * skip('=') * alt_authority > pivot
parameter = ((AltSvcParam << token) * skip('=') *
(token | quoted_string)) > pivot
alt_value = Parametrized << (
alternative *
(MultiDict << many(skip(OWS * ';' * OWS) * parameter))) > pivot
Alt_Svc = clear | comma_list1(alt_value) > pivot
ma = delta_seconds > pivot
persist = subst(True) << literal('1') > pivot
Alt_Used = uri_host + maybe_str(':' + port) > pivot
fill_names(globals(), RFC(7838))
def comma_list1(element):
return _collect_elements << (
many(subst(None) << ',' * OWS) +
((lambda x: [x]) << group(element)) +
many(skip(OWS * ',') * maybe(skip(OWS) * element))
) > named(u'1#rule', RFC(7230, section=(7,)))
def comma_list(element):
return _collect_elements << maybe(
(subst([None, None]) << literal(',') |
(lambda x: [x]) << group(element)) +
many(skip(OWS * ',') * maybe(skip(OWS) * element))) > named(
u'#rule', RFC(7230, section=(7, )))
def media_range(no_q=False):
return Parametrized << (
(literal('*/*') | type_ + '/' + '*' | _check_media_type <<
(MediaType << type_ + '/' + subtype)) * (MultiDict << many(
skip(OWS * ';' * OWS) * parameter(exclude=['q'] if no_q else [])))
) > named(u'media-range', RFC(7231), is_pivot=True)
def transfer_extension(exclude=None, no_q=False):
return Parametrized << (
(TransferCoding << token__excluding(exclude or [])) *
(MultiDict << many(skip(OWS * ';' * OWS) * transfer_parameter(no_q)))
) > named(u'transfer-extension', RFC(7230), is_pivot=True)
':' | '<' | '=' | '>' | '?' | '@' | ALPHA |
'[' | ']' | '^' | '_' | '`' | '{' | '|' |
'}' | '~') > auto
ptoken = string1(ptokenchar) > auto
media_type = MediaType << type_name + '/' + subtype_name > pivot
quoted_mt = skip('"') * media_type * skip('"') > pivot
reg_rel_type = RelationType << (
LOALPHA + string(LOALPHA | DIGIT | '.' | '-')) > auto
ext_rel_type = URI > auto
relation_type = reg_rel_type | ext_rel_type > pivot
relation_types = (
(lambda x: [x]) << relation_type |
skip('"' * OWS) *
(relation_type % many(skip(string1(SP)) * relation_type)) *
skip(OWS * '"')) > pivot
def ext_name_star__excluding(exclude):
return (parmname__excluding(exclude) + '*'
> named(u'ext-name-star', RFC(5988)))
_builtin_params = {
'rel': relation_types,
'anchor': skip('"' * OWS) * URI_Reference * skip(OWS * '"'),
'rev': relation_types,
'hreflang': Language_Tag,
'media': (_MediaDesc_no_delim |
skip('"' * OWS) * _MediaDesc * skip(OWS * '"')),
'title': quoted_string,
'title*': ext_value,
# This has been slightly adapted to the rules of RFC 7230.
# The ``OWS`` are derived from the "implied ``*LWS``" requirement.
# We have no need to special-case "inline" and "attachment", simplify.
disposition_type = CaseInsensitive << token > pivot
filename_parm = (
(CaseInsensitive << literal('filename')) *
skip(OWS * '=' * OWS) * value |
(CaseInsensitive << literal('filename*')) *
skip(OWS * '=' * OWS) * ext_value) > pivot
# ``token`` is a superset of ``ext-token``,
# and special-casing ``ext-token`` requires
# something more complex than our `string_excluding`.
# Until then, we can simplify a bit.
disp_ext_parm = (
(CaseInsensitive << token__excluding(['filename', 'filename*'])) *
skip(OWS * '=' * OWS) * value) > pivot
disposition_parm = filename_parm | disp_ext_parm > auto
content_disposition = Parametrized << (
disposition_type *
(MultiDict << many(skip(OWS * ';' * OWS) * disposition_parm))) > pivot
fill_names(globals(), RFC(6266))
def comma_list(element):
# RFC Errata ID: 5257
return _collect_elements << (
maybe(group(element) * skip(OWS)) %
many(skip(literal(',') * OWS) * maybe(group(element) * skip(OWS)))
) > named(u'#rule', RFC(7230, section=u'7'))
from httpolice.citation import RFC
from httpolice.parse import (
auto,
fill_names,
many,
octet,
octet_range,
pivot,
skip,
string1,
)
from httpolice.syntax.common import SP
from httpolice.syntax.rfc3986 import URI_reference
NQSCHAR = (octet_range(0x20, 0x21) | octet_range(0x23, 0x5B) |
octet_range(0x5D, 0x7E)) > auto
NQCHAR = (octet(0x21) | octet_range(0x23, 0x5B) |
octet_range(0x5D, 0x7E)) > auto
scope_token = string1(NQCHAR) > pivot
scope = scope_token % many(skip(SP) * scope_token) > pivot
error = string1(NQSCHAR) > pivot
error_description = string1(NQSCHAR) > pivot
error_uri = URI_reference > pivot
fill_names(globals(), RFC(6749))
# RFC 7240 Section 2: "Empty or zero-length values on both
# the preference token and within parameters are equivalent
# to no value being specified at all."
(name, value) = x if isinstance(x, tuple) else (x, None)
return Parametrized(name, None if value == u'' else value)
def preference_parameter(head=False):
# The head (first) ``preference-parameter`` of a ``preference``
# contains the actual preference name, which we want to annotate.
name_cls = Preference if head else CaseInsensitive
return (
_normalize_empty_value << (parameter(name_cls=name_cls) |
name_cls << token)
) > named(u'preference-parameter', RFC(7240, errata=4439), is_pivot=True)
preference = Parametrized << (
preference_parameter(head=True) *
many(skip(OWS * ';') * maybe(skip(OWS) * preference_parameter()))
) > named(u'preference', RFC(7240, errata=4439), is_pivot=True)
Prefer = comma_list1(preference) > pivot
Preference_Applied = comma_list1(preference_parameter(head=True)) > pivot
return_ = CaseInsensitive << (literal('representation') | 'minimal') > pivot
wait = delay_seconds > auto
handling = CaseInsensitive << (literal('strict') | 'lenient') > pivot
fill_names(globals(), RFC(7240))
':' | '<' | '=' | '>' | '?' | '@' | ALPHA |
'[' | ']' | '^' | '_' | '`' | '{' | '|' |
'}' | '~') > auto
ptoken = string1(ptokenchar) > auto
media_type = MediaType << type_name + '/' + subtype_name > pivot
quoted_mt = skip('"') * media_type * skip('"') > pivot
reg_rel_type = RelationType << (
LOALPHA + string(LOALPHA | DIGIT | '.' | '-')) > auto
ext_rel_type = URI > auto
relation_type = reg_rel_type | ext_rel_type > pivot
relation_types = (
(lambda x: [x]) << relation_type |
skip('"' * OWS) *
(relation_type % many(skip(string1(SP)) * relation_type)) *
skip(OWS * '"')) > pivot
def ext_name_star__excluding(exclude):
return (parmname__excluding(exclude) + '*'
> named(u'ext-name-star', RFC(5988)))
_builtin_params = {
'rel': relation_types,
'anchor': skip('"' * OWS) * URI_Reference * skip(OWS * '"'),
'rev': relation_types,
'hreflang': Language_Tag,
'media': (_MediaDesc_no_delim |
skip('"' * OWS) * _MediaDesc * skip(OWS * '"')),
'title': quoted_string,
'title*': ext_value,
from httpolice.citation import RFC
from httpolice.parse import (auto, fill_names, many, octet, octet_range, pivot,
skip, string1)
from httpolice.syntax.common import SP
from httpolice.syntax.rfc3986 import URI_reference
NQSCHAR = (octet_range(0x20, 0x21) | octet_range(0x23, 0x5B)
| octet_range(0x5D, 0x7E)) > auto
NQCHAR = (octet(0x21) | octet_range(0x23, 0x5B)
| octet_range(0x5D, 0x7E)) > auto
scope_token = string1(NQCHAR) > pivot
scope = scope_token % many(skip(SP) * scope_token) > pivot
error = string1(NQSCHAR) > pivot
error_description = string1(NQSCHAR) > pivot
error_uri = URI_reference > pivot
fill_names(globals(), RFC(6749))
from httpolice.syntax.rfc2616 import value
from httpolice.syntax.rfc5987 import ext_value
from httpolice.syntax.rfc7230 import OWS, token, token__excluding
# This has been slightly adapted to the rules of RFC 7230.
# The ``OWS`` are derived from the "implied ``*LWS``" requirement.
# We have no need to special-case "inline" and "attachment", simplify.
disposition_type = CaseInsensitive << token > pivot
filename_parm = (
(CaseInsensitive << literal('filename')) * skip(OWS * '=' * OWS) * value |
(CaseInsensitive << literal('filename*')) * skip(OWS * '=' * OWS) *
ext_value) > pivot
# ``token`` is a superset of ``ext-token``,
# and special-casing ``ext-token`` requires
# something more complex than our `string_excluding`.
# Until then, we can simplify a bit.
disp_ext_parm = (
(CaseInsensitive << token__excluding(['filename', 'filename*'])) *
skip(OWS * '=' * OWS) * value) > pivot
disposition_parm = filename_parm | disp_ext_parm > auto
content_disposition = Parametrized << (
disposition_type *
(MultiDict << many(skip(OWS * ';' * OWS) * disposition_parm))) > pivot
fill_names(globals(), RFC(6266))
string_times)
from httpolice.structure import ForwardedParam
from httpolice.syntax.common import ALPHA, DIGIT
from httpolice.syntax.rfc3986 import IPv4address, IPv6address
from httpolice.syntax.rfc7230 import comma_list1, quoted_string, token
def _remove_empty(xs):
return [x for x in xs if x is not None]
obfnode = '_' + string1(ALPHA | DIGIT | '.' | '_' | '-') > pivot
nodename = (IPv4address | skip('[') * IPv6address * skip(']') | 'unknown'
| obfnode) > pivot
port = int << string_times(1, 5, DIGIT) > pivot
obfport = '_' + string1(ALPHA | DIGIT | '.' | '_' | '-') > pivot
node_port = port | obfport > pivot
node = nodename * maybe(skip(':') * node_port) > pivot
value = token | quoted_string > pivot
forwarded_pair = (ForwardedParam << token) * skip('=') * value > pivot
forwarded_element = _remove_empty << (
maybe(forwarded_pair) % many(skip(';') * maybe(forwarded_pair))) > pivot
Forwarded = comma_list1(forwarded_element) > pivot
fill_names(globals(), RFC(7239))
if mtype in _BAD_MEDIA_TYPES:
complain(1282, bad=mtype, good=_BAD_MEDIA_TYPES[mtype])
return mtype
def parameter(exclude=None):
return ((CaseInsensitive << token__excluding(exclude or [])) * skip('=') *
(token | quoted_string)) > named(
u'parameter', RFC(7231), is_pivot=True)
type_ = token > pivot
subtype = token > pivot
media_type = Parametrized << (
(_check_media_type << (MediaType << type_ + '/' + subtype)) *
(MultiDict << many(skip(OWS * ';' * OWS) * parameter()))) > pivot
content_coding = ContentCoding << token > pivot
product_version = token > pivot
product = Versioned << (
(ProductName << token) * maybe(skip('/') * product_version)) > pivot
User_Agent = product % many(
skip(RWS) * (product | comment(include_parens=False))) > pivot
Server = product % many(skip(RWS) *
(product | comment(include_parens=False))) > pivot
day_name = (subst(0) << octet(0x4D) * octet(0x6F) * octet(0x6E)
| subst(1) << octet(0x54) * octet(0x75) * octet(0x65)
| subst(2) << octet(0x57) * octet(0x65) * octet(0x64)
| subst(3) << octet(0x54) * octet(0x68) * octet(0x75)
# which brings in Unicode problems.
_DAY_NAMES = [u'Monday', u'Tuesday', u'Wednesday', u'Thursday', u'Friday',
u'Saturady', u'Sunday']
def parameter(exclude=None):
return (
(CaseInsensitive << token__excluding(exclude or [])) *
skip('=') * (token | quoted_string)
) > named(u'parameter', RFC(7231), is_pivot=True)
type_ = token > pivot
subtype = token > pivot
media_type = Parametrized << (
(MediaType << type_ + '/' + subtype) *
(MultiDict << many(skip(OWS * ';' * OWS) * parameter()))) > pivot
content_coding = ContentCoding << token > pivot
product_version = token > pivot
product = Versioned << ((ProductName << token) *
maybe(skip('/') * product_version)) > pivot
User_Agent = product % many(skip(RWS) *
(product | comment(include_parens=False))) > pivot
Server = product % many(skip(RWS) *
(product | comment(include_parens=False))) > pivot
day_name = (subst(0) << octet(0x4D) * octet(0x6F) * octet(0x6E) |
subst(1) << octet(0x54) * octet(0x75) * octet(0x65) |
subst(2) << octet(0x57) * octet(0x65) * octet(0x64) |
subst(3) << octet(0x54) * octet(0x68) * octet(0x75) |
Transfer_Encoding = comma_list1(transfer_coding()) > pivot
rank = (float << '0' + maybe_str('.' + string_times(0, 3, DIGIT)) |
float << '1' + maybe_str('.' + string_times(0, 3, '0'))) > pivot
t_ranking = skip(OWS * ';' * OWS * 'q=') * rank > pivot
t_codings = (CaseInsensitive << literal('trailers') |
Parametrized << (transfer_coding(no_trailers=True, no_q=True) *
maybe(t_ranking))) > pivot
TE = comma_list(t_codings) > pivot
Trailer = comma_list1(field_name) > pivot
chunk_size = (lambda s: int(s, 16)) << string1(HEXDIG) > pivot
chunk_ext_name = token > auto
chunk_ext_val = token | quoted_string > auto
chunk_ext = many(skip(';') * chunk_ext_name *
maybe(skip('=') * chunk_ext_val)) > pivot
Host = uri_host + maybe_str(':' + port) > pivot
connection_option = ConnectionOption << token > pivot
Connection = comma_list1(connection_option) > pivot
protocol_name = token > pivot
protocol_version = token > pivot
protocol = Versioned << ((UpgradeToken << protocol_name) *
maybe(skip('/') * protocol_version)) > pivot
Upgrade = comma_list1(protocol) > pivot
received_protocol = Versioned << (maybe(protocol_name * skip('/'), u'HTTP') *
protocol_version) > pivot
pseudonym = token > pivot
def _normalize_empty_value(x):
# RFC 7240 Section 2: "Empty or zero-length values on both
# the preference token and within parameters are equivalent
# to no value being specified at all."
(name, value) = x if isinstance(x, tuple) else (x, None)
return Parametrized(name, None if value == u'' else value)
def preference_parameter(head=False):
# The head (first) ``preference-parameter`` of a ``preference``
# contains the actual preference name, which we want to annotate.
name_cls = Preference if head else CaseInsensitive
return (_normalize_empty_value <<
(parameter(name_cls=name_cls) | name_cls << token)) > named(
u'preference-parameter', RFC(7240, errata=4439), is_pivot=True)
preference = Parametrized << (preference_parameter(head=True) * many(
skip(OWS * ';') * maybe(skip(OWS) * preference_parameter()))) > named(
u'preference', RFC(7240, errata=4439), is_pivot=True)
Prefer = comma_list1(preference) > pivot
Preference_Applied = comma_list1(preference_parameter(head=True)) > pivot
return_ = CaseInsensitive << (literal('representation') | 'minimal') > pivot
wait = delay_seconds > auto
handling = CaseInsensitive << (literal('strict') | 'lenient') > pivot
fill_names(globals(), RFC(7240))
def comma_list(element):
return _collect_elements << maybe(
(subst([None, None]) << literal(',') |
(lambda x: [x]) << group(element)) +
many(skip(OWS * ',') * maybe(skip(OWS) * element))
) > named(u'#rule', RFC(7230, section=(7,)))
def comma_list1(element):
return _collect_elements << (many(subst(None) << ',' * OWS) + (
(lambda x: [x]) << group(element)) + many(
skip(OWS * ',') * maybe(skip(OWS) * element))) > named(
u'1#rule', RFC(7230, section=u'7'))
def transfer_extension(exclude=None, no_q=False):
return Parametrized << (
(TransferCoding << token__excluding(exclude or [])) *
(MultiDict << many(skip(OWS * ';' * OWS) * transfer_parameter(no_q)))
) > named(u'transfer-extension', RFC(7230), is_pivot=True)
| float << '1' + maybe_str('.' + string_times(0, 3, '0'))) > pivot
t_ranking = skip(OWS * ';' * OWS * 'q=') * rank > pivot
t_codings = (
CaseInsensitive << literal('trailers') | Parametrized <<
(transfer_coding(no_trailers=True, no_q=True) * maybe(t_ranking))) > pivot
TE = comma_list(t_codings) > pivot
Trailer = comma_list1(field_name) > pivot
chunk_size = (lambda s: int(s, 16)) << string1(HEXDIG) > pivot
chunk_ext_name = token > auto
chunk_ext_val = token | quoted_string > auto
# As updated by RFC 7230 errata ID: 4667.
chunk_ext = many(
skip(BWS * ';' * BWS) * chunk_ext_name *
maybe(skip(BWS * '=' * BWS) * chunk_ext_val)) > pivot
Host = uri_host + maybe_str(':' + port) > pivot
connection_option = ConnectionOption << token > pivot
Connection = comma_list1(connection_option) > pivot
protocol_name = token > pivot
protocol_version = token > pivot
protocol = Versioned << ((UpgradeToken << protocol_name) *
maybe(skip('/') * protocol_version)) > pivot
Upgrade = comma_list1(protocol) > pivot
received_protocol = Versioned << (maybe(protocol_name * skip('/'), u'HTTP') *
protocol_version) > pivot
def comma_list(element):
# RFC Errata ID: 5257
return _collect_elements << (maybe(group(element) * skip(OWS)) % many(
skip(literal(',') * OWS) * maybe(group(element) * skip(OWS)))) > named(
u'#rule', RFC(7230, section=u'7'))
from httpolice.syntax.common import ALPHA, DIGIT
from httpolice.syntax.rfc3986 import IPv4address, IPv6address
from httpolice.syntax.rfc7230 import comma_list1, quoted_string, token
def _remove_empty(xs):
return [x for x in xs if x is not None]
obfnode = '_' + string1(ALPHA | DIGIT | '.' | '_' | '-') > pivot
nodename = (IPv4address |
skip('[') * IPv6address * skip(']') |
'unknown' | obfnode) > pivot
port = int << string_times(1, 5, DIGIT) > pivot
obfport = '_' + string1(ALPHA | DIGIT | '.' | '_' | '-') > pivot
node_port = port | obfport > pivot
node = nodename * maybe(skip(':') * node_port) > pivot
value = token | quoted_string > pivot
forwarded_pair = (ForwardedParam << token) * skip('=') * value > pivot
forwarded_element = _remove_empty << (
maybe(forwarded_pair) % many(skip(';') * maybe(forwarded_pair))) > pivot
Forwarded = comma_list1(forwarded_element) > pivot
fill_names(globals(), RFC(7239))
