3
\9 @ sV d Z ddlZejdjZejdjZejdZejdZejdZ[G dd d Z dS )
zShared support for scanning document type declarations in HTML and XHTML.
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#ddZ d$ddZ dd Z
dd Zdd Zdd Zd d Zd d Zd d! Zd"S )%
ParserBaseziParser base class which provides some common support methods used
by the SGML/HTML and XHTML parsers.c C s | j tkrtdd S )Nz)_markupbase.ParserBase must be subclassed) __class__r RuntimeError)self r #/usr/lib64/python3.6/_markupbase.py__init__ s
zParserBase.__init__c C s t dd S )Nz.subclasses of ParserBase must override error())NotImplementedError)r messager r r error s zParserBase.errorc C s d| _ d| _d S )N r )linenooffset)r r r r reset$ s zParserBase.resetc C s | j | jfS )z&Return current line number and offset.)r
r )r r r r getpos( s zParserBase.getposc C sb ||kr |S | j }|jd||}|rN| j| | _|jd||}||d | _n| j| | | _|S )N
r )rawdatacountr
rindexr )r ijr Znlinesposr r r updatepos0 s zParserBase.updatepos c
C s | j }|d }|||d dkr*|d S |||d dkrBdS t|}|||d dkrh| j|S || dkr~| j|S | j||\}}|dk r|S |d krd| _x||k r|| }|dk r||d | }|d kr| j| n
| j| |d S |d
kr&t||}|s dS |j }n|d krB| j||\} }nt|| jkrX|d }n^|dkr|d kr~| j
|d |}n$|dkr| j d| n
| j d n| j d|| |dk r|S qW dS )N r >-r z--[r Zdoctypez"'Z4abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZattlistlinktypelinkelementz&unsupported '[' char in %s declarationz"unexpected '[' char in declarationz!unexpected %r char in declaration)r r r" > r r r r! r" ) r len
parse_commentparse_marked_section
_scan_name_decl_othercharsZ handle_decl unknown_decl_declstringlit_matchend_parse_doctype_subsetr )
r r r r nZdecltypecdatamnamer r r parse_declaration@ sX
z ParserBase.parse_declarationr c C s | j }| j|d |\}}|dk r&|S |d
kr@tj||d }n4|dkrZtj||d }n| jd ||d | |s|dS |r|jd}| j||d | |jdS )N r tempcdataignoreincludercdataifelseendifz+unknown status keyword %r in marked sectionr > r6 r4 r7 r3 r5 > r8 r9 r: r" ) r r& _markedsectionclosesearch_msmarkedsectioncloser startr( r* )r r reportr ZsectNamer matchr r r r% s
z ParserBase.parse_marked_sectionc C sj | j }|||d dkr$| jd tj||d }|s Zhandle_commentr* )r r r? r r@ r r r r r$ s
zParserBase.parse_commentc
C s4 | j }t|}|}x||k r.|| }|dkr4|||d }|dkrNdS |dkrt| j||d | jd| |d |krdS |d |krdS |||d dkr| j|dd }|dk r|S q| j|d |\}}|dkrdS |dkr | j||d | jd| t| d| } | ||}|dk r,|S q|dkr|d |krPd S | j|d |\}}|dk rr|S || dkr,|d }q|dkr|d }x&||k r|| j r|d }qW ||k r|| dkr|S | j|| | jd nd S q|j r|d }q| j|| | jd| qW d S ) N rD rE r r! r" r" r" )r r# r r r$ r& getattrisspace)
r r declstartposr r, r r- sr0 methr r r r+ 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 rK r0 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 |dkrd||d kr|jd|d }nd S x |||d j r|d }qW ||d sd S n| j||\}}|||d }|s
d
S |dkrLt||}|r.|j }ndS |||d }|sLdS |dkr||d dkrldS | j|d |\}}|dk r|S |||d }|sdS |dkrF|d S qFW d S )Nr r r r ()z'"#r" r" r" r" r" r" r" r" r" )r r& rN rJ r) r* )r r rK r r0 r r- r/ r r r _parse_doctype_attlist sZ
z!ParserBase._parse_doctype_attlistc C s | j ||\}}|dk r |S | j}xl|||d }|s s