�'�X~, � @ s� d d l m Z d d l Z d d l Z d d l Z d d l Z d d l m Z d Z d Z e j
d � Z e j d d k r� e
Z e Z Gd d
� d
e � Z Gd d � d e � Z Gd
d � d e � Z Gd d � d e � Z d d � Z d d � Z d d � Z d d � Z d d � Z d d � Z d d d � Z d d! � Z d" d# � Z d$ d% � Z d& d' � Z d d( d) � Z! d* d+ � Z" d, d- � Z# d. d/ � Z$ d0 d d1 d2 � Z% d d d d d3 d4 � Z& d d d d5 d6 � Z' d S)7� )�idnadata� N)�intranges_contain� s xn--u [.。．｡]� c @ s e Z d Z d Z d S)� IDNAErrorz7 Base exception for all IDNA-encoding related problems N)�__name__�
__module__� __qualname__�__doc__� r r �/core.pyr s r c @ s e Z d Z d Z d S)�
IDNABidiErrorz= Exception when bidirectional requirements are not satisfied N)r r r
r r r r r
r s r c @ s e Z d Z d Z d S)�InvalidCodepointz> Exception when a disallowed or unallocated codepoint is used N)r r r
r r r r r
r s r c @ s e Z d Z d Z d S)�InvalidCodepointContextzE Exception when the codepoint is not valid in the context it is used N)r r r
r r r r r
r s r c C s t j t | � � S)N)� unicodedataZ combining�unichr)�cpr r r
�_combining_class$ s r c C s t t | � t j | � S)N)r �ordr �scripts)r Zscriptr r r
�
_is_script' s r c C s
| j d � S)N�punycode)�encode)�sr r r
� _punycode* s r c C s
d j | � S)Nz U+{0:04X})�format)r r r r
�_unot- s r c C s t | � d k r d Sd S)N�? FT)�len)�labelr r r
�valid_label_length1 s r! c C s& t | � | r d n d k r" d Sd S)N� � FT)r )r � trailing_dotr r r
�valid_string_length8 s r% Fc C s� d } xi t | d � D]X \ } } t j | � } | d k r[ t d j t | � | � � � | d k r d } Pq W| r� | r� d St j | d � } | d k r� d } n0 | d
k r� d } n t d j t | � � � � d } d } x� t | d � D]� \ } } t j | � } | r�| d k r:t d j | � � � | d k rOd } n | d k rad } | d k r�| s|| } q�| | k r�t d � � q� | d k r�t d j | � � � | d k r�d } q� | d k r� d } q� W| s�t d � � d S) NFr � z3Unknown directionality in label {0} at position {1}�R�AL�ANTr �Lz>First codepoint in label {0} must be directionality L, R or AL�EN�ES�CS�ET�ON�BN�NSMzHInvalid direction for codepoint at position {0} in a right-to-left labelz2Can not mix numeral types in a right-to-left labelzHInvalid direction for codepoint at position {0} in a left-to-right labelz0Label ends with illegal codepoint directionality)r' r( r) )r' r( )
r' r( r) r+ r, r- r. r/ r0 r1 )r' r( r+ r) )r) r+ )r* r+ r, r- r. r/ r0 r1 )r* r+ )� enumerater Z
bidirectionalr r �repr) r Z check_ltrZ
bidi_label�idxr � directionZrtlZ valid_endingZ number_typer r r
�
check_bidi? sR
r6 c C s- t j | d � d d k r) t d � � d S)Nr �Mz0Label begins with an illegal combining characterT)r �categoryr )r r r r
�check_initial_combiner| s r9 c C sR | d d � d k r" t d � � | d d k sB | d
d k rN t d � � d S) N� � z--z4Label has disallowed hyphens in 3rd and 4th positionr �-r z)Label must not start or end with a hyphenT���)r )r r r r
�check_hyphen_ok� s
r> c C s( t j d | � | k r$ t d � � d S)N�NFCz%Label must be in Normalization Form C)r � normalizer )r r r r
� check_nfc� s rA c C s� t | | � } | d k rW| d k rL t t | | d � � t k rL d Sd } xt t | d d d � D]\ } t j j t | | � � } | t d � k r� qi | t d � t d � g k ri d } Pqi W| s� d Sd } xw t | d t | � � D]\ } t j j t | | � � } | t d � k r*q� | t d � t d � g k r� d } Pq� W| S| d
k r�| d k r�t t | | d � � t k r�d Sd Sd Sd S)
Ni r r TF�Tr* �Dr' i
r= r= )r r �_virama_combining_class�ranger Z
joining_types�getr )r �pos�cp_value�ok�iZ joining_typer r r
�valid_contextj� s< rK c C s t | | � } | d k r~ d | k o= t | � d k n rz t | | d � d k rz t | | d � d k rz d Sd S| d k r� | t | � d k r� t | � d k r� t | | d d � Sd S| d k s� | d
k r| d k rt | | d d � Sd S| d k rixN | D]F } | d
k r0qt | d � s]t | d � s]t | d � rd SqWd Sd | k o�d k n r�x4 | D], } d t | � k o�d k n r�d Sq�Wd Sd | k o�d k n rx4 | D], } d t | � k od k n r�d Sq�Wd Sd S)N� r r �l TFiu ZGreeki� i� ZHebrewi�0 u ・ZHiraganaZKatakanaZHani` ii i� i� )r r r )r rG � exceptionrH r r r r
�valid_contexto� s> &4 (
-
"
"rO c C s� t | t t f � r$ | j d � } t | � d k rB t d � � t | � t | � t | � xt | � D]\ } } t
| � } t | t j
d � r� qm qm t | t j
d � r� t | | � srt d j t | � | d t | � � � � qm t | t j
d � rGt | | � srt d j t | � | d t | � � � � qm t d
j t | � | d t | � � � � qm Wt | � d S) Nzutf-8r z Empty LabelZPVALIDZCONTEXTJz-Joiner {0} not allowed at position {1} in {2}r ZCONTEXTOz0Codepoint {0} not allowed at position {1} in {2}z0Codepoint {0} at position {1} of {2} not allowed)�
isinstance�bytes� bytearray�decoder r rA r> r9 r2 r r r Zcodepoint_classesrK r r r r3 rO r r6 )r rG r rH r r r
� check_label� s&
../rT c C s� yg | j d � } y t | � Wn' t k
rI t d j | � � � Yn Xt | � sb t d � � | SWn t k
r{ Yn X| s� t d � � t | � } t | � t | � } t | } t | � s� t d � � | S)N�asciiz$The label {0} is not a valid A-labelzLabel too longzNo Input)
r �ulabelr r r! �UnicodeEncodeError�unicoderT r �_alabel_prefix)r r r r
�alabel s(
rZ c C s� t | t t f � sK y | j d � } Wn t k
rJ t | � | SYn X| j � } | j t � r | t t � d � } n t | � | j
d � S| j
d � } t | � | S)NrU r ) rP rQ rR r rW rT �lower�
startswithrY r rS )r r r r
rV s
rV Tc C s� d d l m } d } yCx,t | � D] \ } } t | � } | | d k rS | n t j | | d f � d } | d } t | � d k r� | d n d }
| d k s� | d
k r� | s� | d k r� | r� |
d k r� | | 7} q& |
d k r/| d k s"| d k r| s"| d
k r/| r/| |
7} q& | d
k r& t � � q& Wt j d | � SWn= t k
r�t d j
t | � | d t | � � � � Yn Xd S)zBRe-map the characters in the string according to UTS46 processing.r )� uts46datar& � �Zr r: N�VrC �3r7 �Ir? z0Codepoint {0} not allowed at position {1} in {2})
r] r2 r �bisectZ bisect_leftr �
IndexErrorr r@ r r r r3 ) Zdomain�
std3_rules� transitionalr] �outputrG �charZ
code_pointZuts46row�statusZ replacementr r r
� uts46_remap3 s0
"
rj c C s" t | t t f � r$ | j d � } | r< t | | | � } d } g } | r` | j d � } n t j | � } x | r� | d r� | d =qr W| s� t d � � | d d k r� | d
=d } x! | D] } | j t | � � q� W| r� | j d � d
j
| � } t | | � s t d � � | S)NrU F�.r z Empty domainr r& T� � .zDomain too longr= r= ) rP rQ rR rS rj �split�_unicode_dots_rer �appendrZ �joinr% ) r �strict�uts46re rf r$ �result�labelsr r r r
r O s0
r c C s� t | t t f � r$ | j d � } | r< t | | d � } d } g } | s` t j | � } n | j d � } x | r� | d r� | d =qr W| s� t d � � | d s� | d
=d } x! | D] } | j t | � � q� W| r� | j d � d j
| � S) NrU Frk r z Empty domainr Tr& r= r= ) rP rQ rR rS rj ro rn r rp rV rq )r rr rs re r$ rt ru r r r r
rS l s*
rS )(r& r rc r �re�sysZ intrangesr rD rY �compilero � version_info�strrX �chrr � UnicodeErrorr r r r r r r r r! r% r6 r9 r> rA rK rO rT rZ rV rj r rS r r r r
�<module> sB = ,)