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 }||| dks&td|||d dkrB|d S |||d dkrZdS t|}|||d dkr| j|S || d kr| j|S | j||\}}|d
k r|S |d krd| _x ||k r|| }|dkr||d | }|d kr| j| n
| j| |d S |d kr@t ||}|s6dS |j
}n|d
kr\| j||\} }nt|| jkrr|d }n^|d kr|d kr| j |d |}n$|dkr| j d| n
| j d n| j d|| |d
k r|S qW dS )N z<!z$unexpected call to parse_declarationr >-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 AssertionErrorlen
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@ sZ
z ParserBase.parse_declarationr c C s | j }|||d dks"td| j|d |\}}|dk rB|S |dkr\tj||d }n4|dkrvtj||d }n| jd
||d | |sdS |r|jd}| j||d | |j dS )N z<![z)unexpected call to parse_marked_section()r tempcdataignoreincludercdataifelseendifz+unknown status keyword %r in marked sectionr > r6 r8 r4 r5 r7 > r; r: 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$| jd tj||d }|s<dS |r`|jd}| j||d | |jdS )N z<!--z"unexpected call to parse_comment()r r r" )r r
_commentcloser= r? Zhandle_commentr+ )r r r@ r rA r r r r r% s
zParserBase.parse_commentc
C s4 | j }t|}|}x||k r.|| }|dkr4|||d }|dkrNdS |dkrt| j||d | jd| |d |krdS |d |krdS |||d dkr| j|dd }|dk r|S q| j|d |\}}|dkrdS |dkr | j||d | jd| t| d| } | ||}|dk r,|S q|dkr|d |krPd S | j|d |\}}|dk rr|S || dkr,|d }q|dkr|d }x&||k r|| j r|d }qW ||k r|| dkr|S | j|| | jd nd S q|j r|d }q| j|| | jd| qW d S ) N<r r z<!z*unexpected char in internal subset (in %r)rB z<!--r )r@ r r! entitynotationz)unknown declaration %r in internal subsetZ_parse_doctype_%;]r z%unexpected char after internal subsetz%unexpected char %r in internal subsetr" r" r" r" r" > r r! rF rE r" r" r" )r r$ r r r% r' getattrisspace)
r r declstartposr r- r r. sr1 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 rL r1 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' rO rK r* r+ )r r rL r r1 r r. r0 r r r _parse_doctype_attlist sZ
z!ParserBase._parse_doctype_attlistc C s | j ||\}}|dk r |S | j}xl|||d }|s<dS |dkrL|d S |dkrpt||}|sfdS |j }q$| j ||\}}|dk r$|S q$W d S )Nr r r z'"r" r" )r' r r* r+ )r r rL r1 r r r. r0 r r r _parse_doctype_notation= s$
z"ParserBase._parse_doctype_notationc C s | j }|||d dkrV|d }x6|||d }|s<dS |j rN|d }q$P q$W n|}| j||\}}|dk rv|S xp| j ||d }|sdS |dkrt||}|r|j }qdS qx|dkr|d S | j||\}}|dk rx|S qxW d S ) Nr rG r z'"r r" r" r" )r rK r' r* r+ )r r rL r r r. r1 r0 r r r _parse_doctype_entityT s8
z ParserBase._parse_doctype_entityc C s | j }t|}||krdS t||}|r\|j }|j }|t| |krLdS |j |j fS | j|| | jd|||d d S )Nr zexpected name token at %r r" )Nr" r" )Nr" ) r r$ _declname_matchgroupstriplowerr+ r r )r r rL r r- r0 rM r1 r r r r' x s
zParserBase._scan_namec C s d S )Nr )r r/ r r r r) s zParserBase.unknown_declN)r )r )__name__
__module__ __qualname____doc__r r r r r r( r2 r& r% r, rP rT rU rV r' r) r r r r r s"
R
C 9$r )
r_ recompilerA rX r* rC r< r> r r r r r <module> s
All system for education purposes only. For more tools: Telegram @jackleet