|
Server : Apache System : Linux iad1-shared-b8-43 6.6.49-grsec-jammy+ #10 SMP Thu Sep 12 23:23:08 UTC 2024 x86_64 User : dh_edsupp ( 6597262) PHP Version : 8.2.26 Disable Function : NONE Directory : /lib/python3/dist-packages/future/backports/html/__pycache__/ |
Upload File : |
o
,�]:M � @ s d Z ddlmZmZmZmZ ddlT ddlmZ ddl Z ddl
Z
e � d�Z e � d�Z
e � d�Ze � d �Ze � d
�Ze � d �Ze � d �Ze � d
�Ze � d�Ze � d�Ze � d�Ze � de j�Ze � de j�Ze � d �Ze � d�ZG dd� de �Z G dd� dej �Z dS )zLA parser for HTML and XHTML.
Backported for python-future from Python 3.3.
� )�absolute_import�division�print_function�unicode_literals)�*)� _markupbaseNz[&<]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][-.a-zA-Z0-9:_]*)(?:\s|/(?!>))*z[a-zA-Z][^
/> ]*zJ\s*([a-zA-Z_][-.:a-zA-Z_0-9]*)(\s*=\s*(\'[^\']*\'|"[^"]*"|[^\s"\'=<>`]*))?z]((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*a�
<[a-zA-Z][-.a-zA-Z0-9:_]* # tag name
(?:\s+ # whitespace before attribute name
(?:[a-zA-Z_][-.:a-zA-Z0-9_]* # attribute name
(?:\s*=\s* # value indicator
(?:'[^']*' # LITA-enclosed value
|\"[^\"]*\" # LIT-enclosed value
|[^'\">\s]+ # bare value
)
)?
)
)*
\s* # trailing whitespace
aF
<[a-zA-Z][-.a-zA-Z0-9:_]* # 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 comma
)?(?:\s|/(?!>))*
)*
)?
\s* # trailing whitespace
z#</\s*([a-zA-Z][-.a-zA-Z0-9:_]*)\s*>c @ s" e Zd ZdZddd�Zdd� ZdS ) �HTMLParseErrorz&Exception raised for all parse errors.�NNc C s& |sJ �|| _ |d | _|d | _d S )Nr � ��msg�lineno�offset)�selfr
�position� r �>/usr/lib/python3/dist-packages/future/backports/html/parser.py�__init__U s
zHTMLParseError.__init__c C s> | j }| jd ur|d| j }| jd ur |d| jd }|S )Nz , at line %dz , column %dr r )r �resultr r r �__str__[ s
zHTMLParseError.__str__N)r
)�__name__�
__module__� __qualname__�__doc__r r r r r r r R s
r c @ s� e Zd ZdZdZd:dd�Zdd� Zdd � Zd
d � Zd d
� Z dZ
dd� Z dd� Z dd� Z
dd� Zdd� Zd;dd�Zd d � Zd d � Zd d!� Zd"d#� Zd$d%� Zd&d'� Zd(d)� Zd*d+� Zd,d-� Zd.d/� Zd0d1� Zd2d3� Z d4d5� Z d6d7� Z d8d9� Z dS )<�
HTMLParsera� Find tags and other markup and call handler functions.
Usage:
p = HTMLParser()
p.feed(data)
...
p.close()
Start tags are handled by calling self.handle_starttag() or
self.handle_startendtag(); end tags by self.handle_endtag(). The
data between tags is passed from the parser to the derived class
by calling self.handle_data() with the data as argument (the data
may be split up in arbitrary chunks). Entity references are
passed by calling self.handle_entityref() with the entity
reference as the argument. Numeric character references are
passed to self.handle_charref() with the string containing the
reference as the argument.
)�script�styleFc C s& |r
t jdtdd� || _| �� dS )z�Initialize and reset this instance.
If strict is set to False (the default) the parser will parse invalid
markup, otherwise it will raise an error. Note that the strict mode
is deprecated.
z The strict mode is deprecated.� )�
stacklevelN)�warnings�warn�DeprecationWarning�strict�reset)r r# r r r r z s � zHTMLParser.__init__c C s( d| _ d| _t| _d| _tj�| � dS )z1Reset this instance. Loses all unprocessed data.� z???N)�rawdata�lasttag�interesting_normal� interesting�
cdata_elemr �
ParserBaser$ �r r r r r$ � s
zHTMLParser.resetc C s | j | | _ | �d� dS )z�Feed data to the parser.
Call this as often as you want, with as little or as much text
as you want (may include '\n').
r N)r& �goahead�r �datar r r �feed� s zHTMLParser.feedc C s | � d� dS )zHandle any buffered data.r N)r- r, r r r �close� s zHTMLParser.closec C s t || �� ���N)r �getpos)r �messager r r �error� s zHTMLParser.errorNc C s | j S )z)Return full source of start tag: '<...>'.)�_HTMLParser__starttag_textr, r r r �get_starttag_text� 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_mode� s
zHTMLParser.set_cdata_modec C s t | _d | _d S r2 )r( r) r* r, r r r �clear_cdata_mode� s
zHTMLParser.clear_cdata_modec
C s� | j }d}t|�}||k �r�| j�||�}|r |�� }n| jr!�n�|}||k r0| �|||� � | �||�}||kr<�ng|j}|d|�r�t �
||�rP| � |�}nG|d|�r[| � |�}n<|d|�rf| �
|�}n1|d|�rq| �|�}n&|d|�r�| jr| �|�}n| �|�}n|d |k r�| �d� |d }n�n |dk r�|s��n| jr�| �d� |�d |d �}|dk r�|�d|d �}|dk r�|d }n|d7 }| �|||� � | �||�}n�|d
|��r$t�
||�}|�r |�� d d � } | �| � |�� }|d
|d ��s|d }| �||�}q d
||d � v �r#| �|dd � � | �|d �}n|d|��r�t�
||�}|�rT|�d�} | �| � |�� }|d
|d ��sM|d }| �||�}q t�
||�}|�r�|�r�|�� ||d � k�r�| j�ru| �d� n||k�r||}| �||d �}n |d |k �r�| �d� | �||d �}nnJ d��||k s|�r�||k �r�| j�s�| �|||� � | �||�}||d � | _ d S )Nr �<�</�<!--�<?�<!r zEOF in middle of constructr �&#r ����;�&z#EOF in middle of entity or char refzinteresting.search() lied)r& �lenr) �search�startr* � handle_data� updatepos�
startswith� starttagopen�match�parse_starttag� parse_endtag�
parse_comment�parse_pir# �parse_declaration�parse_html_declarationr5 �find�charref�group�handle_charref�end� entityref�handle_entityref�
incomplete)
r rZ r& �i�nrO �jrM �k�namer r r r- � s�
�
�\ zHTMLParser.goaheadc C s� | j }|||d � dksJ d��|||d � dkr | �|�S |||d � dkr/| �|�S |||d � �� d krX|�d
|d �}|d krId S | �||d |� � |d S | �|�S )
Nr rC z+unexpected call to parse_html_declaration()� rA � z<![� z <!doctyper rE r )r& rR �parse_marked_sectionr8 rV � handle_decl�parse_bogus_comment)r r^ r& �gtposr r r rU s
z!HTMLParser.parse_html_declarationr c C s` | j }|||d � dv sJ d��|�d|d �}|dkr dS |r,| �||d |� � |d S )Nr )rC r@ z"unexpected call to parse_comment()r rE r )r& rV �handle_comment)r r^ �reportr& �posr r r rh - s z HTMLParser.parse_bogus_commentc C sd | j }|||d � dksJ d��t�||d �}|s dS |�� }| �||d |� � |�� }|S )Nr rB z unexpected call to parse_pi()rE )r& �picloserI rJ � handle_pirZ )r r^ r& rO r` r r r rS 9 s zHTMLParser.parse_pic C s" d | _ | �|�}|dk r|S | j}|||� | _ g }t�||d �}|s(J d��|�� }|�d��� | _}||k r�| j rDt
�||�}nt �||�}|sMnT|�ddd�\} }
} |
s\d } n-| d d� d krn| dd � ks�n | d d� d kr�| dd � kr�n n| dd� } | r�| � | �} |�
| �� | f� |�� }||k s:|||� �� } | d vr�| �� \}
}d
| j v r�|
| j �d
� }
t| j �| j �d
� }n|t| j � }| j r�| �d |||� d d � f � | �|||� � |S | �d
�r�| �||� |S | �||� || jv �r| �|� |S )Nr r z#unexpected call to parse_starttag()r rd �'rE �")r �/>�
z junk characters in start tag: %r� rq )r6 �check_for_whole_start_tagr& �tagfindrO rZ rX r8 r' r# �attrfind�attrfind_tolerant�unescape�append�stripr3 �countrH �rfindr5 rK �endswith�handle_startendtag�handle_starttag�CDATA_CONTENT_ELEMENTSr= )r r^ �endposr& �attrsrO ra �tag�m�attrname�rest� attrvaluerZ r r r r r rP E sd
&(
�
��
�
zHTMLParser.parse_starttagc C s | j }| jr
t�||�}nt�||�}|r}|�� }|||d � }|dkr)|d S |dkrY|�d|�r7|d S |�d|�r?dS | jrO| �||d � | �d� ||krU|S |d S |dkr_dS |d v redS | jrs| �||� | �d
� ||kry|S |d S t d ��) Nr r �/rq r rE zmalformed empty start tagr% z6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZzmalformed start tagzwe should not get here!)
r& r# �locatestarttagendrO �locatestarttagend_tolerantrZ rM rL r5 �AssertionError)r r^ r& r� r` �nextr r r rt ~ s>
z$HTMLParser.check_for_whole_start_tagc C sJ | j }|||d � dksJ d��t�||d �}|s dS |�� }t�||�}|s|| jd ur9| �|||� � |S | jrH| � d|||� f � t
�||d �}|se|||d � dkr`|d S | � |�S |� � �
� }|�d |�� �}| �|� |d S |� d��
� }| jd ur�|| jkr�| �|||� � |S | �|�
� � | �� |S )
Nr r@ z unexpected call to parse_endtagr rE zbad end tag: %rrd z</>r )r& � endendtagrI rZ �
endtagfindrO r* rK r# r5 �tagfind_tolerantrh rX r8 rV �
handle_endtagr> )r r^ r& rO ri � namematch�tagnamer< r r r rQ � s<
zHTMLParser.parse_endtagc C s | � ||� | �|� d S r2 )r r� �r r� r� r r r r~ � s z HTMLParser.handle_startendtagc C � d S r2 r r� r r r r � � zHTMLParser.handle_starttagc C r� r2 r )r r� r r r r� � r� zHTMLParser.handle_endtagc C r� r2 r �r rb r r r rY � r� zHTMLParser.handle_charrefc C r� r2 r r� r r r r\ � r� zHTMLParser.handle_entityrefc C r� r2 r r. r r r rK � r� zHTMLParser.handle_datac C r� r2 r r. r r r rj � r� zHTMLParser.handle_commentc C r� r2 r )r �declr r r rg � r� zHTMLParser.handle_declc C r� r2 r r. r r r rn � r� zHTMLParser.handle_pic C s | j r
| �d|f � d S d S )Nzunknown declaration: %r)r# r5 r. r r r � unknown_decl� s �zHTMLParser.unknown_declc C s" d|vr|S dd� }t �d||�S )NrG c S s� | � � d } z-| d dkr2| dd � } | d dv r&t| dd � �d�d�}nt| �d��}t|�W S W n
ty@ d| Y S w ddlm} | |v rO|| S | �d�rXd | S td
t | ��D ]}| d |� |v ry|| d |� | |d � S q_d | S ) Nr �#r )�x�XrF � rD )�html5rG r )
�groups�int�rstrip�chr�
ValueError� future.backports.html.entitiesr� r} �rangerH )�s�cr� r� r r r �replaceEntities� s,
� �
�z,HTMLParser.unescape.<locals>.replaceEntitiesz&&(#?[xX]?(?:[0-9a-fA-F]+;|\w{1,32};?)))r9 �sub)r r� r� r r r rx � s �zHTMLParser.unescape)F)r ) r r r r r� r r$ r0 r1 r5 r6 r7 r= r> r- rU rh rS rP rt rQ r~ r r� rY r\ rK rj rg rn r� rx r r r r r d s<
h
9+* r ) r �
__future__r r r r �future.builtins�future.backportsr r9 r r: r( r] r[ rW rN rm � commentcloseru r� rv rw �VERBOSEr� r� r� r� � Exceptionr r+ r r r r r �<module> s>
��
��