File: //lib64/python3.9/html/__pycache__/parser.cpython-39.opt-2.pyc
a
�DOg�C � @ s� d dl Z d dlZd dlmZ dgZe �d�Ze �d�Ze �d�Ze �d�Z e �d�Z
e �d �Z e �d
�Z e �d �Z
e �d �Ze �d
e j�Ze �d �Ze �d�ZG dd� dej�ZdS )� N)�unescape�
HTMLParserz[&<]z
&[a-zA-Z#]z%&([a-zA-Z][-.a-zA-Z0-9]*)[^a-zA-Z0-9]z)&#(?:[0-9]+|[xX][0-9a-fA-F]+)[^0-9a-fA-F]z <[a-zA-Z]�>z--\s*>z+([a-zA-Z][^\t\n\r\f />\x00]*)(?:\s|/(?!>))*z]((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*aF
<[a-zA-Z][^\t\n\r\f />\x00]* # tag name
(?:[\s/]* # optional whitespace before attribute name
(?:(?<=['"\s/])[^\s/>][^\s/=>]* # attribute name
(?:\s*=+\s* # value indicator
(?:'[^']*' # LITA-enclosed value
|"[^"]*" # LIT-enclosed value
|(?!['"])[^>\s]* # bare value
)
\s* # possibly followed by a space
)?(?:\s|/(?!>))*
)*
)?
\s* # trailing whitespace
z#</\s*([a-zA-Z][-.a-zA-Z0-9:_]*)\s*>c @ s� e Zd ZdZdd�dd�Zdd� Zdd � Zd
d � Zd Zd
d� Z dd� Z
dd� Z dd� Z dd� Z
d6dd�Zdd� Zd d � Zd d � Zd d!� Zd"d#� Zd$d%� Zd&d'� Zd(d)� Zd*d+� Zd,d-� Zd.d/� Zd0d1� Zd2d3� Zd4d5� Z d S )7r )Zscript�styleT)�convert_charrefsc C s || _ | �� d S �N)r �reset)�selfr � r
�#/usr/lib64/python3.9/html/parser.py�__init__V s zHTMLParser.__init__c C s( d| _ d| _t| _d | _tj�| � d S )N� z???)�rawdata�lasttag�interesting_normal� interesting�
cdata_elem� _markupbase�
ParserBaser �r r
r
r r _ s
zHTMLParser.resetc C s | j | | _ | �d� d S )Nr )r �goahead�r �datar
r
r �feedg s zHTMLParser.feedc C s | � d� d S )N� )r r r
r
r �closep s zHTMLParser.closeNc C s | j S r )�_HTMLParser__starttag_textr r
r
r �get_starttag_textv s z HTMLParser.get_starttag_textc C s$ |� � | _t�d| j tj�| _d S )Nz </\s*%s\s*>)�lowerr �re�compile�Ir )r �elemr
r
r �set_cdata_modez s
zHTMLParser.set_cdata_modec C s t | _d | _d S r )r r r r r
r
r �clear_cdata_mode~ s zHTMLParser.clear_cdata_modec C sJ | j }d}t|�}||k �r�| jrv| jsv|�d|�}|dk r�|�dt||d ��}|dkrpt�d�� ||�sp�q�|}n*| j
� ||�}|r�|� � }n| jr��q�|}||k r�| jr�| js�| � t
|||� �� n| � |||� � | �||�}||kr��q�|j}|d|��rJt�||��r"| �|�} n�|d|��r:| �|�} nn|d|��rR| �|�} nV|d|��rj| �|�} n>|d |��r�| �|�} n&|d
|k �r�| � d� |d
} n�q�| dk �r<|�s��q�|�d |d
�} | dk �r�|�d|d
�} | dk �r|d
} n| d
7 } | j�r*| j�s*| � t
||| � �� n| � ||| � � | �|| �}q|d |��r�t�||�}|�r�|�� d
d� }
| �|
� |�� } |d| d
��s�| d
} | �|| �}qn<d||d � v �r�| � |||d
� � | �||d
�}�q�q|d|�rt�||�}|�rN|�d
�}
| � |
� |�� } |d| d
��s@| d
} | �|| �}qt �||�}|�r�|�r�|�� ||d � k�r�|�� } | |k�r�|} | �||d
�}�q�n.|d
|k �r�| � d� | �||d
�}n�q�qq|�r8||k �r8| j�s8| j�r| j�s| � t
|||� �� n| � |||� � | �||�}||d � | _ d S )Nr �<�&�" z[\s;]z</�<!--z<?z<!r r z&#� ����;) r �lenr r �find�rfind�maxr r �searchr �start� handle_datar Z updatepos�
startswith� starttagopen�match�parse_starttag� parse_endtag�
parse_comment�parse_pi�parse_html_declaration�charref�group�handle_charref�end� entityref�handle_entityref�
incomplete) r r> r �i�n�jZampposr5 r3 �k�namer
r
r r � s�
�
zHTMLParser.goaheadc C s� | j }|||d � dkr$| �|�S |||d � dkrB| �|�S |||d � �� dkr�|�d|d �}|dkrvdS | �||d |� � |d
S | �|�S d S ) N� r( � z<![� z <!doctyper r* r) r )r r8 Zparse_marked_sectionr r- � handle_decl�parse_bogus_comment)r rB r �gtposr
r
r r: � s
z!HTMLParser.parse_html_declarationr c C sD | j }|�d|d �}|dkr"dS |r<| �||d |� � |d S )Nr r) r* r )r r- �handle_comment)r rB �reportr �posr
r
r rK s z HTMLParser.parse_bogus_commentc C sH | j }t�||d �}|s dS |�� }| �||d |� � |�� }|S )Nr) r* )r �picloser0 r1 � handle_pir> )r rB r r5 rD r
r
r r9 s zHTMLParser.parse_pic C s� d | _ | �|�}|dk r |S | j}|||� | _ g }t�||d �}|�� }|�d��� | _}||k �r t �||�}|s~�q |�ddd�\} }
} |
s�d } nZ| d d� d kr�| dd � ks�n | d d� d kr�| dd � kr�n n | dd� } | �rt
| �} |� | �� | f� |�� }q`|||� � � } | dv�r�| �
� \}
}d | j v �rz|
| j �d � }
t| j �| j �d � }n|t| j � }| �|||� � |S | �d
��r�| �||� n"| �||� || jv �r�| �|� |S ) Nr r r) rH �'r* �")r �/>�
rT )r �check_for_whole_start_tagr �tagfind_tolerantr5 r> r< r r �attrfind_tolerantr �append�stripZgetpos�countr, r. r2 �endswith�handle_startendtag�handle_starttag�CDATA_CONTENT_ELEMENTSr# )r rB �endposr �attrsr5 rE �tag�m�attrname�restZ attrvaluer> �lineno�offsetr
r
r r6 , sX
& �
�
�
zHTMLParser.parse_starttagc C s� | j }t�||�}|r�|�� }|||d � }|dkr>|d S |dkr~|�d|�rZ|d S |�d|�rjdS ||krv|S |d S |dkr�dS |dv r�dS ||kr�|S |d S td ��d S )
Nr r �/rT r) r* r
z6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZzwe should not get here!)r �locatestarttagend_tolerantr5 r> r3 �AssertionError)r rB r rc rD �nextr
r
r rV _ s. z$HTMLParser.check_for_whole_start_tagc C s | j }t�||d �}|s dS |�� }t�||�}|s�| jd urV| �|||� � |S t�||d �}|s�|||d � dkr�|d S | � |�S |�
d�� � }|� d|�� �}| �
|� |d S |�
d�� � }| jd ur�|| jkr�| �|||� � |S | �
|� | �� |S )Nr r* r) rH z</>r )r � endendtagr0 r> �
endtagfindr5 r r2 rW rK r< r r- �
handle_endtagr$ )r rB r r5 rL Z namematchZtagnamer" r
r
r r7 � s6
zHTMLParser.parse_endtagc C s | � ||� | �|� d S r )r^ rn �r rb ra r
r
r r] � s z HTMLParser.handle_startendtagc C s d S r r
ro r
r
r r^ � s zHTMLParser.handle_starttagc C s d S r r
)r rb r
r
r rn � s zHTMLParser.handle_endtagc C s d S r r
�r rF r
r
r r= � s zHTMLParser.handle_charrefc C s d S r r
rp r
r
r r@ � s zHTMLParser.handle_entityrefc C s d S r r
r r
r
r r2 � s zHTMLParser.handle_datac C s d S r r
r r
r
r rM � s zHTMLParser.handle_commentc C s d S r r
)r Zdeclr
r
r rJ � s zHTMLParser.handle_declc C s d S r r
r r
r
r rQ � s zHTMLParser.handle_pic C s d S r r
r r
r
r � unknown_decl� s zHTMLParser.unknown_decl)r ) �__name__�
__module__� __qualname__r_ r r r r r r r# r$ r r: rK r9 r6 rV r7 r] r^ rn r= r@ r2 rM rJ rQ rq r
r
r
r r > s4 z
3"()r r Zhtmlr �__all__r r rA r? r; r4 rP Z commentcloserW rX �VERBOSEri rl rm r r r
r
r
r �<module> s(
��