3
\9 � @ sV d Z ddlZejd�jZejd�jZejd�Zejd�Zejd�Z[G dd � d �Z dS )
z�Shared support for scanning document type declarations in HTML and XHTML.
This module is used as a foundation for the html.parser module. It has no
documented public API and should not be used directly.
� Nz[a-zA-Z][-_.a-zA-Z0-9]*\s*z(\'[^\']*\'|"[^"]*")\s*z--\s*>z ]\s*]\s*>z]\s*>c @ s� e Zd ZdZdd� Zdd� Zdd� Zdd � Zd
d � Zd Z d
d� Z
d#dd�Z d$dd�Z dd� Z
dd� Zdd� Zdd� Zd d � Zd d � Zd d!� Zd"S )%�
ParserBaseziParser base class which provides some common support methods used
by the SGML/HTML and XHTML parsers.c C s | j tkrtd��d S )Nz)_markupbase.ParserBase must be subclassed)� __class__r � RuntimeError)�self� r �#/usr/lib64/python3.6/_markupbase.py�__init__ s
zParserBase.__init__c C s t d��d S )Nz.subclasses of ParserBase must override error())�NotImplementedError)r �messager r r �error s zParserBase.errorc C s d| _ d| _d S )N� r )�lineno�offset)r r r r �reset$ s zParserBase.resetc C s | j | jfS )z&Return current line number and offset.)r
r )r r r r �getpos( s zParserBase.getposc C sb ||kr |S | j }|jd||�}|rN| j| | _|jd||�}||d | _n| j| | | _|S )N�
r )�rawdata�countr
�rindexr )r �i�jr Znlines�posr r r � updatepos0 s zParserBase.updatepos� c
C s� | j }|d }|||d � dkr*|d S |||d � dkrBdS t|�}|||d � dkrh| j|�S || dkr~| j|�S | j||�\}}|dk r�|S |d kr�d| _�x||k �r�|| }|dk� r�||d |� }|d kr�| j|� n
| j|� |d S |d
k�r&t||�}|�s dS |j � }n�|d k�rB| j||�\} }nt|| jk�rX|d }n^|dk�r�|d k�r~| j
|d |�}n$|dk�r�| j d| � n
| j d� n| j d|| � |dk r�|S q�W dS )N� r �>�-r z--�[r Zdoctypez"'Z4abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ�attlist�linktype�link�elementz&unsupported '[' char in %s declarationz"unexpected '[' char in declarationz!unexpected %r char in declaration)r r ���r" > r r! r r r" ) r �len�
parse_comment�parse_marked_section�
_scan_name�_decl_othercharsZ handle_decl� unknown_decl�_declstringlit_match�end�_parse_doctype_subsetr )
r r r r �nZdecltype�c�data�m�namer r r �parse_declaration@ sX
z ParserBase.parse_declarationr c C s� | j }| j|d |�\}}|dk r&|S |d
kr@tj||d �}n4|dkrZtj||d �}n| jd ||d |� � |s|dS |r�|jd�}| j||d |� � |jd�S )N� r �temp�cdata�ignore�include�rcdata�if�else�endifz+unknown status keyword %r in marked sectionr > r4 r5 r6 r3 r7 > r9 r8 r: r" ) r r& �_markedsectionclose�search�_msmarkedsectioncloser �startr( r* )r r �reportr ZsectNamer �matchr r r r% � s
z ParserBase.parse_marked_sectionc C sj | j }|||d � dkr$| jd� tj||d �}|s<dS |r`|jd�}| j||d |� � |jd�S )N� z<!--z"unexpected call to parse_comment()r r r" )r r �
_commentcloser< r> Zhandle_commentr* )r r r? r r@ r r r r r$ � s
zParserBase.parse_commentc
C s4 | j }t|�}|}�x||k �r.|| }|dk�r4|||d � }|dkrNdS |dkrt| j||d � | jd| � |d |kr�dS |d |kr�dS |||d � dkr�| j|dd �}|dk r|S q| j|d |�\}}|dkr�dS |dk�r | j||d � | jd| � t| d| �} | ||�}|dk �r,|S q|dk�r�|d |k�rPd S | j|d |�\}}|dk �rr|S || dk�r,|d }q|dk�r�|d }x&||k �r�|| j� �r�|d }�q�W ||k �r�|| dk�r�|S | j||� | jd� nd S q|j� �r|d }q| j||� | jd| � qW d S ) N�<r r z<!z*unexpected char in internal subset (in %r)rA z<!--r )r? r r! �entity�notationz)unknown declaration %r in internal subsetZ_parse_doctype_�%�;�]r z%unexpected char after internal subsetz%unexpected char %r in internal subsetr" r" r" r" r" > rD r! r rE r" r" r" )r r# r r r$ r& �getattr�isspace)
r r � declstartposr r, r r- �sr0 �methr r r r+ � sn
z ParserBase._parse_doctype_subsetc C sF | j ||�\}}|dkr dS | j}d||d � krB|jd|�d S dS )Nr r r" r" r" )r& r �find)r r rK r0 r r r r r �_parse_doctype_element� s z!ParserBase._parse_doctype_elementc C s� | j }| j||�\}}|||d � }|dkr2d S |dkrB|d S �xv| j||�\}}|dk rb|S |||d � }|dkr~d
S |dkr�d||d � kr�|jd|�d }nd S x |||d � j� r�|d }q�W ||d � s�d S n| j||�\}}|||d � }|�s
d
S |dk�rLt||�}|�r.|j� }ndS |||d � }|�sLdS |dk�r�||d � dk�rldS | j|d |�\}}|dk �r�|S |||d � }|�s�dS |dkrF|d S qFW d S )Nr r r r �(�)z'"�#r" r" r" r" r" r" r" r" r" )r r& rN rJ r) r* )r r rK r r0 r r- r/ r r r �_parse_doctype_attlist sZ
z!ParserBase._parse_doctype_attlistc C s� | j ||�\}}|dk r |S | j}xl|||d � }|s<dS |dkrL|d S |dkrpt||�}|sfdS |j� }q$| j ||�\}}|dk r$|S q$W d S )Nr r r z'"r" r" )r& r r) r* )r r rK r0 r r r- r/ r r r �_parse_doctype_notation= s$
z"ParserBase._parse_doctype_notationc C s� | j }|||d � dkrV|d }x6|||d � }|s<dS |j� rN|d }q$P q$W n|}| j||�\}}|dk rv|S xp| j ||d � }|s�dS |dkr�t||�}|r�|j� }q�dS qx|dkr�|d S | j||�\}}|dk rx|S qxW d S ) Nr rF r z'"r r" r" r" )r rJ r& r) r* )r r rK r r r- r0 r/ r r r �_parse_doctype_entityT s8
z ParserBase._parse_doctype_entityc C s� | j }t|�}||krdS t||�}|r\|j� }|j� }|t|� |krLdS |j� |j� fS | j||� | jd|||d � � d S )Nr zexpected name token at %r� r" )Nr" r" )Nr" ) r r# �_declname_match�group�strip�lowerr* r r )r r rK r r, r/ rL r0 r r r r& x s
zParserBase._scan_namec C s d S )Nr )r r. r r r r( � s zParserBase.unknown_declN)r )r )�__name__�
__module__� __qualname__�__doc__r r r r r r' r1 r% r$ r+ rO rS rT rU r& r( r r r r r s"
R
C 9$r )
r^ �re�compiler@ rW r) rB r; r= r r r r r �<module> s