U
    ,º]W?  ã                   @   sX   d Z ddlZe d¡jZe d¡jZe d¡Ze d¡Ze d¡Z[G dd	„ d	e	ƒZ
dS )
a  Shared support for scanning document type declarations in HTML and XHTML.

Backported for python-future from Python 3.3. Reason: ParserBase is an
old-style class in the Python 2.7 source of markupbase.py, which I suspect
might be the cause of sporadic unit-test failures on travis-ci.org with
test_htmlparser.py.  The test failures look like this:

    ======================================================================

ERROR: test_attr_entity_replacement (future.tests.test_htmlparser.AttributesStrictTestCase)

----------------------------------------------------------------------

Traceback (most recent call last):
  File "/home/travis/build/edschofield/python-future/future/tests/test_htmlparser.py", line 661, in test_attr_entity_replacement
    [("starttag", "a", [("b", "&><"'")])])
  File "/home/travis/build/edschofield/python-future/future/tests/test_htmlparser.py", line 93, in _run_check
    collector = self.get_collector()
  File "/home/travis/build/edschofield/python-future/future/tests/test_htmlparser.py", line 617, in get_collector
    return EventCollector(strict=True)
  File "/home/travis/build/edschofield/python-future/future/tests/test_htmlparser.py", line 27, in __init__
    html.parser.HTMLParser.__init__(self, *args, **kw)
  File "/home/travis/build/edschofield/python-future/future/backports/html/parser.py", line 135, in __init__
    self.reset()
  File "/home/travis/build/edschofield/python-future/future/backports/html/parser.py", line 143, in reset
    _markupbase.ParserBase.reset(self)

TypeError: unbound method reset() must be called with ParserBase instance as first argument (got EventCollector instance instead)

This module is used as a foundation for the html.parser module.  It has no
documented public API and should not be used directly.

é    Nz[a-zA-Z][-_.a-zA-Z0-9]*\s*z(\'[^\']*\'|"[^"]*")\s*z--\s*>z	]\s*]\s*>z]\s*>c                   @   s   e Zd ZdZdd„ Zdd„ Zdd„ Zdd	„ Zd
d„ ZdZ	dd„ Z
d#dd„Zd$dd„Zdd„ Zdd„ Zdd„ Zdd„ Zdd„ Zdd„ Zd d!„ Zd"S )%Ú
ParserBaseziParser base class which provides some common support methods used
    by the SGML/HTML and XHTML parsers.c                 C   s   | j tkrtdƒ‚d S )Nz)_markupbase.ParserBase must be subclassed)Ú	__class__r   ÚRuntimeError©Úself© r   ú>/usr/lib/python3/dist-packages/future/backports/_markupbase.pyÚ__init__6   s    
ÿzParserBase.__init__c                 C   s   t dƒ‚d S )Nz.subclasses of ParserBase must override error())ÚNotImplementedError)r   Úmessager   r   r   Úerror;   s    ÿzParserBase.errorc                 C   s   d| _ d| _d S )Né   r   ©ÚlinenoÚoffsetr   r   r   r   Úreset?   s    zParserBase.resetc                 C   s   | j | jfS )z&Return current line number and offset.r   r   r   r   r   ÚgetposC   s    zParserBase.getposc                 C   sb   ||kr|S | j }| d||¡}|rN| j| | _| d||¡}||d  | _n| j| | | _|S )NÚ
r   )ÚrawdataÚcountr   Úrindexr   )r   ÚiÚjr   ZnlinesÚposr   r   r   Ú	updateposK   s    zParserBase.updateposÚ c           
      C   sê  | j }|d }|||… dks&tdƒ‚|||d … dkrB|d S |||d … dkrZdS t|ƒ}|||d … dkr€|  |¡S || d	kr–|  |¡S |  ||¡\}}|d
k r²|S |dkrÀd| _||k ræ|| }|dkr||d |… }|dkr|  |¡ n
|  |¡ |d S |dkr<t	||ƒ}|s2dS | 
¡ }nœ|dkrX|  ||¡\}	}n€|| jkrn|d }nj|d	krÆ|dkr”|  |d |¡}n0|tddddgƒkrº|  d| ¡ n
|  d¡ n|  d||  ¡ |d
k rÀ|S qÀdS )Né   ú<!z$unexpected call to parse_declarationr   ú>)ú-r   éÿÿÿÿz--ú[r   Zdoctyper   z"'Z4abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZÚattlistZlinktypeÚlinkÚelementz&unsupported '[' char in %s declarationz"unexpected '[' char in declarationz!unexpected %r char in declaration)r   ÚAssertionErrorÚlenÚparse_commentÚparse_marked_sectionÚ
_scan_nameÚ_decl_othercharsZhandle_declÚunknown_declÚ_declstringlit_matchÚendÚ_parse_doctype_subsetÚsetr   )
r   r   r   r   ÚnZdecltypeÚcÚdataÚmÚnamer   r   r   Úparse_declaration[   s\    













ÿzParserBase.parse_declarationr   c                 C   sÞ   | j }|||d … dks"tdƒ‚|  |d |¡\}}|dk rB|S |tddddd	gƒkrjt ||d ¡}n>|td
ddgƒkrŽt ||d ¡}n|  d||d |…  ¡ |s°dS |rÔ| d¡}|  	||d |… ¡ | 
d¡S )Né   z<![z)unexpected call to parse_marked_section()r   ZtempZcdataÚignoreZincludeZrcdataÚifÚelseZendifz+unknown status keyword %r in marked sectionr    )r   r%   r)   r/   Ú_markedsectioncloseÚsearchÚ_msmarkedsectioncloser   Ústartr+   r-   )r   r   Úreportr   ZsectNamer   Úmatchr   r   r   r(   ­   s     
zParserBase.parse_marked_sectionc                 C   sj   | j }|||d … dkr$|  d¡ t ||d ¡}|s<dS |r`| d¡}|  ||d |… ¡ | d¡S )Né   ú<!--z"unexpected call to parse_comment()r    r   )r   r   Ú_commentcloser;   r=   Zhandle_commentr-   )r   r   r>   r   r?   r   r   r   r   r'   Ã   s    

zParserBase.parse_commentc           
      C   s6  | j }t|ƒ}|}||k r2|| }|dkr<|||d … }|dkrJdS |dkrp|  ||d ¡ |  d| ¡ |d |kr€dS |d |krdS |||d … dkrÀ| j|d	d
}|d	k r|S q|  |d |¡\}}|dkràdS |tddddgƒkr|  ||d ¡ |  d| ¡ t| d| ƒ}	|	||ƒ}|d	k r0|S q|dkr’|d |krXdS |  |d |¡\}}|d	k rz|S || dkr0|d }q|dkr|d }||k rÈ||  ¡ rÈ|d }q¤||k rü|| dkrä|S |  ||¡ |  d¡ ndS q| ¡ r|d }q|  ||¡ |  d| ¡ qdS )Nú<r   r    r   r   z*unexpected char in internal subset (in %r)r@   rA   r   )r>   r"   r$   ZentityZnotationz)unknown declaration %r in internal subsetZ_parse_doctype_ú%ú;ú]r   z%unexpected char after internal subsetz%unexpected char %r in internal subset)	r   r&   r   r   r'   r)   r/   ÚgetattrÚisspace)
r   r   Údeclstartposr   r0   r   r1   Úsr4   Úmethr   r   r   r.   Ñ   sp    

ÿ








z ParserBase._parse_doctype_subsetc                 C   sF   |   ||¡\}}|dkrdS | j}d||d … krB| d|¡d S dS )Nr    r   r   )r)   r   Úfind)r   r   rI   r4   r   r   r   r   r   Ú_parse_doctype_element  s    z!ParserBase._parse_doctype_elementc                 C   s¶  | j }|  ||¡\}}|||d … }|dkr2dS |dkrB|d S |  ||¡\}}|dk r^|S |||d … }|dkrzdS |dkrØd||d … kr¤| d|¡d }ndS |||d …  ¡ rÆ|d }q¨||d … sèdS n|  ||¡\}}|||d … }|sdS |dkrDt||ƒ}|r&| ¡ }ndS |||d … }|sDdS |d	kr ||d … d	krddS |  |d |¡\}}|dk r†|S |||d … }|s dS |dkrB|d S qBd S )
Nr   r   r    r   r   ú(ú)ú'"ú#)r   r)   rL   rH   r,   r-   )r   r   rI   r   r4   r   r1   r3   r   r   r   Ú_parse_doctype_attlist  sX    





z!ParserBase._parse_doctype_attlistc                 C   s   |   ||¡\}}|dk r|S | j}|||d … }|s:dS |dkrJ|d S |dkrnt||ƒ}|sddS | ¡ }q"|   ||¡\}}|dk r"|S q"d S )Nr   r   r    r   rP   )r)   r   r,   r-   )r   r   rI   r4   r   r   r1   r3   r   r   r   Ú_parse_doctype_notationX  s"    

z"ParserBase._parse_doctype_notationc                 C   sä   | j }|||d … dkrR|d }|||d … }|s:dS | ¡ rP|d }q"qVq"n|}|  ||¡\}}|dk rr|S | j ||d … }|sŒdS |dkr²t||ƒ}|r¬| ¡ }qÞdS qr|dkrÂ|d S |  ||¡\}}|dk rr|S qrd S )Nr   rD   r    r   rP   r   )r   rH   r)   r,   r-   )r   r   rI   r   r   r1   r4   r3   r   r   r   Ú_parse_doctype_entityo  s4    


z ParserBase._parse_doctype_entityc                 C   s†   | j }t|ƒ}||krdS t||ƒ}|r\| ¡ }| ¡ }|t|ƒ |krLdS | ¡ | ¡ fS |  ||¡ |  d|||d …  ¡ d S )N)Nr    zexpected name token at %ré   )	r   r&   Ú_declname_matchÚgroupÚstripÚlowerr-   r   r   )r   r   rI   r   r0   r3   rJ   r4   r   r   r   r)   “  s    
ÿzParserBase._scan_namec                 C   s   d S )Nr   )r   r2   r   r   r   r+   ¥  s    zParserBase.unknown_declN)r   )r   )Ú__name__Ú
__module__Ú__qualname__Ú__doc__r	   r   r   r   r   r*   r5   r(   r'   r.   rM   rR   rS   rT   r)   r+   r   r   r   r   r   2   s"   R

C9$r   )r]   ÚreÚcompiler?   rV   r,   rB   r:   r<   Úobjectr   r   r   r   r   Ú<module>   s   "


