/snap/core18/2999/usr/lib/python3.6/html/__pycache__
Edit: /snap/core18/2999/usr/lib/python3.6/html/__pycache__/parser.cpython-36.pyc (11595B)
3
Nq i I @ s d Z ddlZddlZddlZddlmZ dgZejdZejdZ ejdZ
ejdZ ejd Z ejd
Z
ejd Zejd Zejd
ZejdZejdejZejd ZejdZG dd dejZdS )z A parser for HTML and XHTML. 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
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 comma
)?(?:\s|/(?!>))*
)*
)?
\s* # trailing whitespace
z#c @ s e Zd ZdZd:ZddddZdd Zd
d Zd d
ZdZ dd Z
dd Z dd Z dd Z
dd Zd;ddZd d Zd d Zd d! Zd"d# Zd$d% Zd&d' Zd(d) Zd*d+ Zd,d- Zd.d/ Zd0d1 Zd2d3 Zd4d5 Z d6d7 Z d8d9 Z dS )
'.)_HTMLParser__starttag_text)r r
r
r get_starttag_textx s z HTMLParser.get_starttag_textc C s$ |j | _tjd| j tj| _d S )Nz )lowerr recompileIr )r elemr
r
r set_cdata_mode| s
zHTMLParser.set_cdata_modec C s t | _d | _d S )N)r r r )r r
r
r clear_cdata_mode s zHTMLParser.clear_cdata_modec C sB | j }d}t|}x||k r| jr|| j r||jd|}|dk r|jdt||d }|dkrvtjdj || rvP |}n(| j
j ||}|r|j }n | jrP |}||k r| jr| j r| j t
||| n| j ||| | j||}||krP |j}|d|r4tj||r&| j|} n|d|r>| j|} nr|d|rV| j|} nZ|d|rn| j|} nB|d |r| j|} n*|d
|k s|r| j d |d
} nP | dk r&|sP tj||rԐnN|d|r$|d |kr| j d n&tj||r
n| j||d d n|d|r||}x.dD ]&}
|j|
|d r:|t|
8 }P q:W | j||d | n|d|r| j||d d n|||d j dkr| j ||d d nP|d |r| j||d d n,|d|r| j ||d d nt d|} | j|| }q|d|rt j||}|r|j d d } | j!| |j" } |d| d
s| d
} | j|| }qn:d||d kr| j |||d | j||d }P q|d|rt#j||}|r8|j d
} | j$| |j" } |d| d
s*| d
} | j|| }qt%j||}|r|r|j ||d kr|j" } | |kr||} | j||d
}P n,|d
|k r| j d | j||d
}nP qdst dqW |r0||k r0| j r0| jr| j r| j t
||| n| j ||| | j||}||d | _ d S )Nr <&" z[\s;]zd S |rX| j||d | |d S )
Nr&
r. r. r. )r rZ )r check_for_whole_start_tagr tagfind_tolerantr8 rD rH rF r r attrfind_tolerantr appendstripZgetposcountr/ r1 r5 r@ handle_startendtaghandle_starttagCDATA_CONTENT_ELEMENTSr! )r rL endposr attrsr8 rO tagmZattrnamerestZ attrvaluerH linenooffsetr
r
r r9 @ sR
(*
zHTMLParser.parse_starttagc C s | j }tj||}|r|j }|||d }|dkr>|d S |dkr~|jd|rZ|d S |jd|rjd S ||krv|S |d S |dkrd
S |dkrd S ||kr|S |d S tdd S ) Nr r /z/>r& r
z6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZzwe should not get here!r. r. r. )r locatestarttagend_tolerantr8 rH r6 rD )r rL r rh rN nextr
r
r r\ s s. z$HTMLParser.check_for_whole_start_tagc C s2 | j }|||d dks"tdtj||d }|s:dS |j }tj||}|s| jd k rr| j||| |S t j||d }|s|||d dkr|d S | j
|S |j dj }|j
d|j }| j| |d S |j dj }| jd k r|| jkr| j||| |S | j|j | j |S ) Nr& zr r. )r rD endendtagr3 rH
endtagfindr8 r r5 r] rR rF r r0
handle_endtagr" )r rL r r8 rS Z namematchZtagnamer r
r
r r: s8
zHTMLParser.parse_endtagc C s | j || | j| d S )N)rc rq )r rg rf r
r
r rb s z HTMLParser.handle_startendtagc C s d S )Nr
)r rg rf r
r
r rc s zHTMLParser.handle_starttagc C s d S )Nr
)r rg r
r
r rq s zHTMLParser.handle_endtagc C s d S )Nr
)r rQ r
r
r rG s zHTMLParser.handle_charrefc C s d S )Nr
)r rQ r
r
r rJ s zHTMLParser.handle_entityrefc C s d S )Nr
)r r r
r
r r5 s zHTMLParser.handle_datac C s d S )Nr
)r r r
r
r r? s zHTMLParser.handle_commentc C s d S )Nr
)r Zdeclr
r
r rB s zHTMLParser.handle_declc C s d S )Nr
)r r r
r
r rC s zHTMLParser.handle_pic C s d S )Nr
)r r r
r
r rA s zHTMLParser.unknown_declc C s t jdtdd t|S )NzZThe unescape method is deprecated and will be removed in 3.5, use html.unescape() instead.r& )
stacklevel)warningswarnDeprecationWarningr )r sr
r
r r s
zHTMLParser.unescape)r r )r ) __name__
__module__ __qualname____doc__rd r r r r r r r! r" r r= rR r< r9 r\ r: rb rc rq rG rJ r5 r? rB rC rA r r
r
r
r r @ s:
3"()rz r rs r Zhtmlr __all__r r rK rI rE r7 r> rW Z commentcloser] r^ VERBOSErm ro rp r r r
r
r
r s*