ruạṛ
3 AL�iW? � @ sX d Z ddlZejd�jZejd�jZejd�Zejd�Zejd�Z[G dd � d e �Z dS ) a Shared support for scanning document type declarations in HTML and XHTML. Backported for python-future from Python 3.3. Reason: ParserBase is an old-style class in the Python 2.7 source of markupbase.py, which I suspect might be the cause of sporadic unit-test failures on travis-ci.org with test_htmlparser.py. The test failures look like this: ====================================================================== ERROR: test_attr_entity_replacement (future.tests.test_htmlparser.AttributesStrictTestCase) ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/travis/build/edschofield/python-future/future/tests/test_htmlparser.py", line 661, in test_attr_entity_replacement [("starttag", "a", [("b", "&><"'")])]) File "/home/travis/build/edschofield/python-future/future/tests/test_htmlparser.py", line 93, in _run_check collector = self.get_collector() File "/home/travis/build/edschofield/python-future/future/tests/test_htmlparser.py", line 617, in get_collector return EventCollector(strict=True) File "/home/travis/build/edschofield/python-future/future/tests/test_htmlparser.py", line 27, in __init__ html.parser.HTMLParser.__init__(self, *args, **kw) File "/home/travis/build/edschofield/python-future/future/backports/html/parser.py", line 135, in __init__ self.reset() File "/home/travis/build/edschofield/python-future/future/backports/html/parser.py", line 143, in reset _markupbase.ParserBase.reset(self) TypeError: unbound method reset() must be called with ParserBase instance as first argument (got EventCollector instance instead) 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$dd�Z dd� Z dd� Zdd� Zdd� Zd d � Zd d � Zd d!� Zd"S )%� ParserBaseziParser base class which provides some common support methods used by the SGML/HTML and XHTML parsers.c C s | j tkrtd��d S )Nz)_markupbase.ParserBase must be subclassed)� __class__r � RuntimeError)�self� r �C/root/tmp/pip-build-3x0y2zol/future/future/backports/_markupbase.py�__init__6 s zParserBase.__init__c C s t d��d S )Nz.subclasses of ParserBase must override error())�NotImplementedError)r �messager r r �error; s zParserBase.errorc C s d| _ d| _d S )N� r )�lineno�offset)r r r r �reset? s zParserBase.resetc C s | j | jfS )z&Return current line number and offset.)r r )r r r r �getposC s zParserBase.getposc C sb ||kr |S | j }|jd||�}|rN| j| | _|jd||�}||d | _n| j| | | _|S )N� r )�rawdata�countr �rindexr )r �i�jr �nlines�posr r r � updateposK s zParserBase.updatepos� c C s� | j }|d }|||� dks&td��|||d � dkrB|d S |||d � dkrZdS t|�}|||d � dkr�| j|�S || d kr�| j|�S | j||�\}}|d k r�|S |d kr�d| _�x(||k �r�|| }|dk�r||d |� }|d k�r| j|� n | j|� |d S |d k�r@t ||�}|�s6dS |j � }n�|d k�r\| j||�\} }n�|| jk�rr|d }nj|d k�r�|d k�r�| j |d |�}n0|t ddddg�k�r�| j d| � n | j d� n| j d|| � |d k r�|S q�W dS )N� z<!z$unexpected call to parse_declarationr �>�-r z--�[r �doctypez"'�4abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ�attlist�linktype�link�elementz&unsupported '[' char in %s declarationz"unexpected '[' char in declarationz!unexpected %r char in declaration)r r ���r% r% )r �AssertionError�len� parse_comment�parse_marked_section� _scan_name�_decl_otherchars� handle_decl� unknown_decl�_declstringlit_match�end�_parse_doctype_subset�setr ) r r r r �n�decltype�c�data�m�namer r r �parse_declaration[ sZ z ParserBase.parse_declarationr c C s� | j }|||d � dks"td��| j|d |�\}}|dk rB|S |tddddd g�krjtj||d �}n>|td d d g�kr�tj||d �}n| jd ||d |� � |s�dS |r�|jd�}| j ||d |� � |j d�S )N� z<![z)unexpected call to parse_marked_section()r �temp�cdata�ignore�include�rcdata�if�else�endifz+unknown status keyword %r in marked sectionr r% ) r r& r* r1 �_markedsectionclose�search�_msmarkedsectioncloser �startr- r/ )r r �reportr �sectNamer �matchr r r r) � s z ParserBase.parse_marked_sectionc C sj | j }|||d � dkr$| jd� tj||d �}|s<dS |r`|jd�}| j||d |� � |jd�S )N� z<!--z"unexpected call to parse_comment()r r r% )r r � _commentcloserC rE �handle_commentr/ )r r rF r rH r r r r r( � s zParserBase.parse_commentc C s@ | j }t|�}|}�x&||k �r:|| }|dk�r@|||d � }|dkrNdS |dkrt| j||d � | jd| � |d |kr�dS |d |kr�dS |||d � dkr�| j|dd �}|dk r|S q| j|d |�\}}|dkr�dS |td d d d g�k�r| j||d � | jd| � t| d| �} | ||�}|dk �r8|S q|dk�r�|d |k�r\dS | j|d |�\}}|dk �r~|S || dk�r8|d }q|dk�r |d }x&||k �r�|| j� �r�|d }�q�W ||k �r|| dk�r�|S | j||� | jd� nd S q|j� �r |d }q| j||� | jd| � qW d S ) N�<r r z<!z*unexpected char in internal subset (in %r)rI z<!--r )rF r! r$ �entity�notationz)unknown declaration %r in internal subset�_parse_doctype_�%�;�]r z%unexpected char after internal subsetz%unexpected char %r in internal subsetr% r% r% r% r% r% r% r% ) r r' r r r( r* r1 �getattr�isspace) r r � declstartposr r2 r r4 �sr7 �methr r r r0 � 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 rU r7 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* rX rT r. r/ )r r rU r r7 r r4 r6 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$W d S )Nr r r z'"r% r% )r* r r. r/ )r r rU r7 r r r4 r6 r r r �_parse_doctype_notationX s$ 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 rP r z'"r r% r% r% )r rT r* r. r/ )r r rU r r r4 r7 r6 r r r �_parse_doctype_entityo 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 rU r r2 r6 rV r7 r r r r* � s zParserBase._scan_namec C s d S )Nr )r r5 r r r r- � s zParserBase.unknown_declN)r )r )�__name__� __module__� __qualname__�__doc__r r r r r r+ r8 r) r( r0 rY r] r^ r_ r* r- r r r r r 2 s" R C 9$r ) rh �re�compilerH ra r. rJ rB rD �objectr r r r r �<module>! s
cải xoăn