| Linux sh00002.bigrock.com 5.14.0-687.29.1.el9_8.x86_64 #1 SMP PREEMPT_DYNAMIC Thu Jul 23 16:18:48 EDT 2026 x86_64 Path : /lib64/python3.9/html/__pycache__/ |
| Current File : //lib64/python3.9/html/__pycache__/parser.cpython-39.opt-2.pyc |
a
�|jrW � @ 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
�Ze �de j�Ze �de j�Ze �de 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
</[a-zA-Z]�>z--!?>z-?>z0([a-zA-Z][^\t\n\r\f />]*)(?:[\t\n\r\f ]|/(?!>))*a{
(
(?<=['"\t\n\r\f /])[^\t\n\r\f />][^\t\n\r\f /=>]* # attribute name
)
([\t\n\r\f ]*=[\t\n\r\f ]* # value indicator
('[^']*' # LITA-enclosed value
|"[^"]*" # LIT-enclosed value
|(?!['"])[^>\t\n\r\f ]* # bare value
)
)?
(?:[\t\n\r\f ]|/(?!>))* # possibly followed by a space
a
[a-zA-Z][^\t\n\r\f />]* # tag name
[\t\n\r\f /]* # optional whitespace before attribute name
(?:(?<=['"\t\n\r\f /])[^\t\n\r\f />][^\t\n\r\f /=>]* # attribute name
(?:[\t\n\r\f ]*=[\t\n\r\f ]* # value indicator
(?:'[^']*' # LITA-enclosed value
|"[^"]*" # LIT-enclosed value
|(?!['"])[^>\t\n\r\f ]* # bare value
)
)?
[\t\n\r\f /]* # possibly followed by a space
)*
>?
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 ZdZdZddd�dd�Zdd � Zd
d � Zd d
� ZdZ dd� Z
dd�dd�Z dd� Z d=dd�Z
dd� Zdd� Zd>d d �Zd?d d �Zd!d"� Zd#d$� Zd%d&� Zd'd(� Zd)d*� Zd+d,� Zd-d.� Zd/d0� Zd1d2� Zd3d4� Zd5d6� Z d7d8� Z d9d:� Z d;d<� Z dS )@r )Zscript�styleZxmpZiframeZnoembedZnoframes)Ztextarea�titleTF)�convert_charrefs� scriptingc C s || _ || _| �� d S �N)r r �reset)�selfr r � r �#/usr/lib64/python3.9/html/parser.py�__init__v s
zHTMLParser.__init__c C sF d| _ d| _t| _d | _d| _d| _g | _d| _d| _ t
j � | � d S )N� z???Tr � )
�rawdata�lasttag�interesting_normal� interesting�
cdata_elem�_support_cdata�
_escapable�_pending� _pending_len�_parse_threshold� _markupbase�
ParserBaser
�r r r r
r
� s zHTMLParser.resetc C s� | j t|�7 _ | j | jk r,| j�|� n~| jsB| j|7 _n,| j�|� | jd�| j�7 _| j�� d| _ t| j�}| �d� t| j�|k r�d| _n t| j�| _d S �Nr r r ) r �lenr r �appendr �join�clear�goahead)r �data�nr r r
�feed� s
zHTMLParser.feedc C s: | j r,| jd�| j �7 _| j �� d| _| �d� d S r )r r r! r" r r# r r r r
�close� s
zHTMLParser.closeNc C s | j S r )�_HTMLParser__starttag_textr r r r
�get_starttag_text� s z HTMLParser.get_starttag_text�� escapablec C sp |� � | _|| _| jdkr(t�d�| _nD|rP| jsPt�d| j tjtjB �| _n t�d| j tjtjB �| _d S )N� plaintextz\Zz&|</%s(?=[\t\n\r\f />])z</%s(?=[\t\n\r\f />])) �lowerr r �re�compiler r �
IGNORECASE�ASCII)r �elemr+ r r r
�set_cdata_mode� s
�
�zHTMLParser.set_cdata_modec C s t | _d | _d| _d S )NT)r r r r r r r r
�clear_cdata_mode� s zHTMLParser.clear_cdata_modec C s
|| _ d S r )r )r �flagr r r
�_set_support_cdata� s z HTMLParser._set_support_cdatac C s4 | j }d}t|�}||k �r�| jrv| jsv|�d|�}|dk r�|�dt||d ��}|dkrpt�d�� ||�sp�q�|}n*| j
� ||�}|r�|� � }n| jr��q�|}||k r�| jr�| j r�| �
t|||� �� n| �
|||� � | �||�}||kr��q�|j}|d|��r<t�||��r"| �|�} n�|d|��r:| �|�} nt|d|��rR| �|�} n\|d|��rj| �|�} nD|d |��r�| �|�} n,|d
|k �s�|�r�| �
d� |d
} n�q�| dk �r.|�sq�t�||��rԐnV|d|��r$|d |k�r�| �
d� n&t�||��r
n| �||d d � � �n|d|��r||}d D ]*}
|�|
|d
��r8|t|
�8 } �qd�q8| �||d
|� � n�|d|��r�| j�r�| � ||d d � � n�|||d � � � dk�r�| � ||d d � � nP|d |��r�| �||d d � � n,|d|��r"| � ||d d � � nt d��|} | �|| �}q|d|��r�t!�||�}|�r�|�"� d d� } | �#| � |�$� } |d| d
��s�| d
} | �|| �}qn<d||d � v �r�| �
|||d � � | �||d �}�q�q|d|�rt%�||�}|�r@|�"d
�} | �&| � |�$� } |d| d
��s2| d
} | �|| �}qt'�||�}|�r�|�r�|�"� ||d � k�r�|�$� } | |k�r�|} | �||d
�}�q�n.|d
|k �r�| �
d� | �||d
�}n�q�qq|�r"||k �r"| j�r| j �r| �
t|||� �� n| �
|||� � | �||�}||d � | _ d S )Nr �<�&�" z [\t\n\r\f ;]z</�<!--z<?z<!r � )z--!z--�-� � <![CDATA[� � � <!doctypezwe should not get here!z&#����;)(r r r r �find�rfind�maxr. r/ �searchr �startr � handle_datar Z updatepos�
startswith� starttagopen�match�parse_starttag� parse_endtag�
parse_comment�parse_pi�parse_html_declaration�
endtagopen�handle_comment�endswithr � unknown_declr- � handle_decl� handle_pi�AssertionError�charref�group�handle_charref�end� entityref�handle_entityref�
incomplete) r r\ r �ir% �jZampposrL rJ �k�suffix�namer r r
r# � s�
�
zHTMLParser.goaheadc C sX | j }|||d � dkr$| �|�S |||d � dkrx| jrx|�d|d �}|dk rZdS | �||d |� � |d S |||d � �� d kr�|�d
|d �}|dkr�dS | �||d |� � |d S |||d � d
k�rJ|�d
|d �}|dk � r�dS ||d dk�r,| �||d |d � � n| �||d |� � |d S | �|�S d S )Nr= r: r@ r> z]]>r rB r? rA r r; r z<![�]) r rO r rD rU r- rV rS �parse_bogus_comment)r r` r ra Zgtposr r r
rQ j s0
z!HTMLParser.parse_html_declarationc C s\ | j }t�||d �}|s2t�||d �}|s2dS |rT|�� }| �||d |� � |�� S )Nr= rB )r � commentcloserG �commentabruptcloserL rH rS r\ )r r` �reportr rL ra r r r
rO � s zHTMLParser.parse_commentr c C sD | j }|�d|d �}|dkr"dS |r<| �||d |� � |d S )Nr r; rB r )r rD rS )r r` ri r �posr r r
rf � s z HTMLParser.parse_bogus_commentc C sH | j }t�||d �}|s dS |�� }| �||d |� � |�� }|S )Nr; rB )r �picloserG rH rW r\ �r r` r rL ra r r r
rP � 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 �s�| j�r�|d k�s�|d k�r�| j|d
d� n|| jv �r| j|dd� |S )Nr r r; r? �'rB �")r �/>�
ro Znoscriptr, Fr* T)r( �check_for_whole_start_tagr �tagfind_tolerantrL r\ rZ r- r �attrfind_tolerantr r �stripZgetpos�countr rE rI rT �handle_startendtag�handle_starttag�CDATA_CONTENT_ELEMENTSr r3 �RCDATA_CONTENT_ELEMENTS)r r` �endposr �attrsrL rb �tag�m�attrname�restZ attrvaluer\ �lineno�offsetr r r
rM � sh
& �
�
� ��� zHTMLParser.parse_starttagc C s6 | j }t�||d �}|�� }||d dkr2dS |S )Nr r rB )r � locatetagendrL r\ rl r r r
rq � s z$HTMLParser.check_for_whole_start_tagc C s� | j }|�d|d �dk r dS t�||�sT||d |d � dkrJ|d S | �|�S t�||d �}|�� }||d dkr�dS t�||d �}|�d�� � }| �
|� | � � |S )Nr r; r rB r? r ) r rD rR rL rf r� r\ rr rZ r- �
handle_endtagr4 )r r` r rL ra r| r r r
rN � s
zHTMLParser.parse_endtagc C s | � ||� | �|� d S r )rw r� �r r| r{ r r r
rv s z HTMLParser.handle_startendtagc C s d S r r r� r r r
rw s zHTMLParser.handle_starttagc C s d S r r )r r| r r r
r� s zHTMLParser.handle_endtagc C s d S r r �r rd r r r
r[ ! s zHTMLParser.handle_charrefc C s d S r r r� r r r
r^ % s zHTMLParser.handle_entityrefc C s d S r r �r r$ r r r
rI ) s zHTMLParser.handle_datac C s d S r r r� r r r
rS - s zHTMLParser.handle_commentc C s d S r r )r Zdeclr r r
rV 1 s zHTMLParser.handle_declc C s d S r r r� r r r
rW 5 s zHTMLParser.handle_pic C s d S r r r� r r r
rU 8 s zHTMLParser.unknown_decl)T)T)r ) �__name__�
__module__� __qualname__rx ry r r
r& r' r( r) r3 r4 r6 r# rQ rO rf rP rM rq rN rv rw r� r[ r^ rI rS rV rW rU r r r r
r Z s<
"
9
)r. r Zhtmlr �__all__r/ r r_ r] rY rK rR rk rg rh rr �VERBOSErs r� Zlocatestarttagend_tolerantZ endendtagZ
endtagfindr r r r r r
�<module> s2
�
��