�
�A�[c@`sz
dZd
dl
mZmZmZmZ
d
dlTd
dlmZ
d
dl	Z	d
dl
Z
e	jd�
Ze	jd�
Z
e	jd�
Ze	jd	�
Ze	jd
�
Ze	jd�
Ze	jd�
Ze	jd
�
Ze	jd�
Ze	jd�
Ze	jd�
Ze	jde	j�Ze	jde	j�Ze	jd�
Ze	jd�
Zdef
d��YZdejf
d��YZdS(uLA parser for HTML and XHTML.

Backported for python-future from Python 3.3.
i(tabsolute_importtdivisiontprint_functiontunicode_literals(
t
*(
t_markupbaseNu[&<]u
&[a-zA-Z#]u%&([a-zA-Z][-.a-zA-Z0-9]*)[^a-zA-Z0-9]u)&#(?:[0-9]+|[xX][0-9a-fA-F]+)[^0-9a-fA-F]u	<[a-zA-Z]u
>u--\s*>u(([a-zA-Z][-.a-zA-Z0-9:_]*)(?:\s|/(?!>))*u[a-zA-Z][^	

 />]*uJ\s*([a-zA-Z_][-.:a-zA-Z_0-9]*)(\s*=\s*(\'[^\']*\'|"[^"]*"|[^\s"\'=<>`]*))?u]((?<=[\'"\s/])[^\s/>][^\s/=>]*)(\s*=+\s*(\'[^\']*\'|"[^"]*"|(?![\'"])[^>\s]*))?(?:\s|/(?!>))*u�

  <[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
uF
  <[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
u#</\s*([a-zA-Z][-.a-zA-Z0-9:_]*)\s*>tHTMLParseErrorcB`s#eZ
dZdd
�
Zd�ZRS(u&Exception raised for all parse errors.cC`s3|
st�
|
|_
|d
|_|d|_dS(Nii
(tAssertionErrortmsgtlinenotoffset(tselfRtposition((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyt__init__Us

	


c
C`sW|j}
|j
dk	r,|
d
|j
}
n|jdk	rS|
d|jd}
n|
S(Nu, at line %du, column %di
(RR	tNoneR
(Rtresult((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyt__str__[s
	




N(NN(t__name__t
__module__t__doc__RR
R(((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRRs
t
HTMLParsercB`s

eZ
dZdZed�
Zd�Zd�Zd�Zd�Z	dZd�Zd	�Z
d
�Zd�Zd�Zd
d�
Zd�Zd�Zd�Zd�Zd�Zd�Zd�Zd�Zd�Zd�Zd�Zd�Zd�Zd�Zd�Z RS( u�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.
    uscriptustylecC`s6|
rtj
d
tdd�

n|
|_|j�
dS(u�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.
        uThe strict mode is deprecated.t
stackleveliN(twarningstwarntDeprecationWarningtstricttreset(RR((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR
zs

	

	
c

C`s8d
|_d|_
t|_d|_tjj|�

dS(u1Reset this instance.  Loses all unprocessed data.uu???N(	trawdatatlasttagtinteresting_normaltinterestingRt
cdata_elemRt
ParserBaseR(
R((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR�s
	
	
	
	
cC`s!|j|
|_|j
d
�

dS(u�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').
        iN(Rtgoahead(Rtdata((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pytfeed�s
c

C`s|jd
�

dS(uHandle any buffered data.i
N(
R!(
R((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pytclose�scC`st|
|j
���
dS(
N(Rtgetpos(Rtmessage((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyterror�s
c


C`s|jS(
u)Return full source of start tag: '<...>'.(
t_HTMLParser__starttag_text(
R((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pytget_starttag_text�scC`s2|
j�|_
tjd
|j
tj�|_dS(Nu</\s*%s\s*>(tlowerRtretcompilet
IR(Rtelem((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pytset_cdata_mode�s

c

C`st|_
d|_dS(
N(RRRR(
R((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pytclear_cdata_mode�s
	
c
C`s�|j}d
}t
|�
}xk||kr�|jj||�}|rT|j�}n|jraPn|}||kr�|j|||!�

n|j||�}||kr�Pn|j}|d|�rkt	j
||�r�|j|�
}n�|d|�r	
|j|�
}n�|d|�r*
|j
|�
}n�|d|�rK
|j|�
}ng|d|�r�
|jru
|j|�
}q�
|j|�
}n+|d|kr�
|jd�

|d}n
P|d
krV|
s�
Pn|jr�
|jd�

n|jd	|d�}|d
kr5|jd|d�}|d
kr?|d}q?n
|d7}|j|||!�

n|j||�}q|d
|�r3tj
||�}|r�|j�dd!}	|j|	�

|j�}|d
|d�s�|d}n|j||�}qq�d
||kr/|j|d
d!�

|j|d�}nPq|d|�rstj
||�}|r�|jd�
}	|j|	�

|j�}|d
|d�s�|d}n|j||�}qntj
||�}|r9|
r5|j�||kr5|jr|jd�

q5||k
r|}n|j||d�}nPq�|d|kro|jd�

|j||d�}q�Pqd
std�
�
qW|
r�||kr�|jr�|j|||!�

|j||�}n|||_dS(Niu
<u</u<!--u<?u<!i
uEOF in middle of constructu
>u&#ii����u
;u
&u#EOF in middle of entity or char refuinteresting.search() lied(RtlenRtsearchtstartRthandle_datat	updatepost
startswithtstarttagopentmatchtparse_starttagtparse_endtagt
parse_commenttparse_piRtparse_declarationtparse_html_declarationR'tfindtcharreftgroupthandle_charreftendt	entityrefthandle_entityreft
incompleteR(
RRCRt
it
nR8t
jR6t
ktname((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR!�s�
	





	





	










	










	






































	

	






cC`s�|j}||
|
d
!dks,t
d�
�
||
|
d!dkrP|j|
�
S||
|
d!dkrt|j|
�
S||
|
d!j�d	kr�|jd
|
d�}|dkr�dS|j||
d
|!�

|dS|j|
�
SdS(
Niu<!u+unexpected call to parse_html_declaration()iu<!--iu<![i	u	<!doctypeu
>i����i
(RRR;tparse_marked_sectionR*R?thandle_decltparse_bogus_comment(RRGRtgtpos((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR>
s
	
#








i
cC`s{|j}||
|
d
!dks,t
d�
�
|jd|
d
�}|dkrRdS|rs|j||
d
|!�

n|dS(	Niu<!u</u"unexpected call to parse_comment()u
>i����i
(u<!u</(RRR?thandle_comment(RRGtreportRtpos((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRN-
s
	
#




cC`s�|j}||
|
d
!dks,t
d�
�
tj||
d
�}|sLdS|j�}|j||
d
|!�

|j�}|S(Niu<?uunexpected call to parse_pi()i����(RRtpicloseR2R3t	handle_piRC(RRGRR8RI((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR<9
s
	
#






cC`s�d|_
|j|
�
}|d
kr(|S|j}||
|!|_
g}tj||
d�}|sotd�
�
|j�}|jd�
j	�|_
}x
||kr�
|jr�tj||�}nt
j||�}|s�Pn|jddd�\}	}
}|
s

d}nX|d dko,
|dkn
sU
|d dkoP
|dkn
re
|dd!}n|r}
|j|�
}n|j|	j	�|f�

|j�}q�W|||!j�}|dkrt|j�\}
}d|j
kr|
|j
jd�
}
t|j
�
|j
jd�
}n|t|j
�
}|jr\|jd|||!d
 f
�

n|j||
|!�

|S|jd
�
r�|j||�
n/|j||�
||jkr�|j|�

n|S(Nii
u#unexpected call to parse_starttag()iiu
'i����u
"u
>u/>u

u junk characters in start tag: %ri(u
>u/>(RR(tcheck_for_whole_start_tagRttagfindR8RRCRAR*RRtattrfindtattrfind_toleranttunescapetappendtstripR%tcountR1trfindR'R4tendswiththandle_startendtagthandle_starttagtCDATA_CONTENT_ELEMENTSR/(RRGtendposRtattrsR8RJttagt
mtattrnametrestt	attrvalueRCR	R
((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR9E
s\
	



	






	





	
$
$











	
	






cC`sh
|j}|j
r'tj||
�}ntj||
�}|rX
|j�}|||d
!}|dkrp|d
S|dkr�|jd|�r�|dS|jd|�r�dS|j
r�|j|
|d
�
|jd�

n||
kr�|S|
d
Sn|dkr
dS|d	kr
dS|j
r=
|j|
|�
|jd
�

n||
krM
|S|
d
Snt	d�
�
dS(Ni
u
>u
/u/>ii����umalformed empty start taguu6abcdefghijklmnopqrstuvwxyz=/ABCDEFGHIJKLMNOPQRSTUVWXYZumalformed start taguwe should not get here!(
RRtlocatestarttagendR8tlocatestarttagend_tolerantRCR6R5R'R(RRGRReRItnext((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRU~
s>
	
	









	






	




cC`s�
|j}||
|
d
!dks,t
d�
�
tj||
d�}|sLdS|j�}tj||
�}|sH
|jdk	r�|j	||
|!�

|S|j
r�|jd||
|!f
�

ntj||
d
�}|s	
||
|
d!dkr�|
dS|j
|
�
Sn|j�j�}|jd	|j��}|j|�

|dS|jd�
j�}|jdk	r�
||jkr�
|j	||
|!�

|Sn|j|j��

|j�
|S(
Niu</uunexpected call to parse_endtagi
i����ubad end tag: %riu</>u
>(RRt	endendtagR2RCt
endtagfindR8RRR4RR'ttagfind_tolerantRNRAR*R?t
handle_endtagR0(RRGRR8ROt	namematchttagnameR.((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR:�
s<
	
#









	













cC`s!|j|
|�
|j
|
�

dS(
N(R`Ro(RRdRc((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR_�
s

c
C`sdS(
N((RRdRc((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR`�
s
c
C`sdS(
N((RRd((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRo�
s
c
C`sdS(
N((RRK((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRB�
s
c
C`sdS(
N((RRK((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRE�
s
c
C`sdS(
N((RR"((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyR4�
s
c
C`sdS(
N((RR"((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRP�
s
c
C`sdS(
N((Rtdecl((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRM�
s
c
C`sdS(
N((RR"((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRT�
s
cC`s$|jr |j
d
|
f
�

ndS(Nuunknown declaration: %r(RR'(RR"((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pytunknown_decl�
s
	
cC`s,d
|
kr|
Sd�}tj
d||
�S(Nu
&c
S`s 
|j�d
}yl|d
dkr{|d}|d
dkr\t
|djd�
d�}
nt
|jd�
�
}
t|
�
SWntk
r�


d|SXd
d	lm}
||kr�||S|jd�
r�d
|SxItdt	|�
�D]*}|| |kr�||| ||Sq�Wd
|SdS(
Niu
#i
u
xu
Xu
;iu&#(
thtml5u
&i(u
xu
X(
tgroupstinttrstriptchrt
ValueErrortfuture.backports.html.entitiesRtR^trangeR1(t
st
cRtt
x((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pytreplaceEntities�
s&










	






u&&(#?[xX]?(?:[0-9a-fA-F]+;|\w{1,32};?))(R+tsub(RR|R((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRY�
s



		
(uscriptustyleN(!RRRRatFalseR
RR#R$R'RR(R)R/R0R!R>RNR<R9RUR:R_R`RoRBRER4RPRMRTRsRY(((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyRds:
									h			9	+	*										( Rt
__future__RR
RRtfuture.builtinstfuture.backportsRR+RR,RRFRDR@R7RStcommentcloseRVRnRWRXtVERBOSERiRjRlRmt	ExceptionRR R(((sM/opt/alt/python27/lib/python2.7/site-packages/future/backports/html/parser.pyt<module>s4	"