o
aia~ @ s d dl mZmZmZ d dlmZ d dlmZmZ d dl Z d dl
Z
d dl m Z m
Z
d dlmZ ddlmZmZmZmZ dd lmZ dd
lmZ ed d eD Zed
d eD Zedd eD ZeeddgB Z dZ ej re d dkr{e ddks}J e
e dd e!d d Z"ne
e Z"h dZ#e
dZ$i Z%G dd de&Z'dd Z(G dd d e&Z)G d d d e)Z*G d d d e+Z,G d!d" d"e&Z-G d#d$ d$e&Z.d%d& Z/dS )' )absolute_importdivisionunicode_literals) text_type) http_clienturllibN)BytesIOStringIO) webencodings )EOFspaceCharacters asciiLettersasciiUppercase)_ReparseException)_utilsc C g | ]}| d qS asciiencode.0item r /builddir/build/BUILDROOT/alt-python310-pip-21.3.1-5.el8.x86_64/opt/alt/python310/lib/python3.10/site-packages/pip/_vendor/html5lib/_inputstream.py
r c C r r r r r r r r r c C r r r r r r r r r > i
i i i i i i i i i i i i i i i i i i i
i i i i i i i
i
i i i i z[ -
-/:-@\[-`{-~]c @ sH e Zd ZdZdd Zdd Zdd Zdd Zd
d Zd d
Z dd Z
dS )BufferedStreamzBuffering for streams that do not have buffering of their own
The buffer is implemented as a list of chunks on the assumption that
joining many strings will be slow since it is O(n**2)
c C s || _ g | _ddg| _d S )Nr r )streambufferposition)selfr# r r r __init__: s zBufferedStream.__init__c C s< d}| j d | jd D ]}|t|7 }q || jd 7 }|S Nr r )r$ r% len)r& poschunkr r r tell? s
zBufferedStream.tellc C sd || ksJ |}d}t| j| |k r+|t| j| 8 }|d7 }t| j| |k s||g| _d S r( )_bufferedBytesr) r$ r% )r& r* offsetir r r seekF s zBufferedStream.seekc C sP | j s| |S | jd t| j kr#| jd t| j d kr#| |S | |S )Nr r r )r$ _readStreamr% r) _readFromBufferr& bytesr r r readO s
zBufferedStream.readc C s t dd | jD S )Nc S s g | ]}t |qS r )r) r r r r r Y s z1BufferedStream._bufferedBytes..)sumr$ r& r r r r- X z BufferedStream._bufferedBytesc C s< | j |}| j| | jd d7 < t|| jd< |S r( )r# r5 r$ appendr% r) )r& r4 datar r r r1 [ s
zBufferedStream._readStreamc C s |}g }| j d }| j d }|t| jk rc|dkrc|dks J | j| }|t|| kr6|}||| g| _ nt|| }|t|g| _ |d7 }||||| ||8 }d}|t| jk rc|dks|rm|| | d|S )Nr r )r% r) r$ r9 r1 join)r& r4 remainingBytesrv bufferIndex bufferOffset bufferedData bytesToReadr r r r2 b s(
z BufferedStream._readFromBufferN) __name__
__module__ __qualname____doc__r' r, r0 r5 r- r1 r2 r r r r r" 3 s r" c K s t | tjst | tjjrt | jtjrd}nt| dr%t | dt }nt | t }|rCdd |D }|r;t
d| t | fi |S t | fi |S )NFr5 r c S s g | ] }| d r|qS ) _encoding)endswith)r xr r r r s z#HTMLInputStream..z3Cannot set an encoding with a unicode input, set %r)
isinstancer HTTPResponser responseaddbasefphasattrr5 r TypeErrorHTMLUnicodeInputStreamHTMLBinaryInputStream)sourcekwargs isUnicode encodingsr r r HTMLInputStream} s
rW c @ sp e Zd ZdZdZdd Zdd Zdd Zd d
Zd d Z d
d Z
dddZ dd Z dd Z
d ddZdd ZdS ) rQ Provides a unicode stream of characters to the HTMLTokenizer.
This class takes care of character encoding and removing or replacing
incorrect byte-sequences and also provides column and line tracking.
i ( c C sZ t jsd| _ntddkr| j| _n| j| _dg| _tddf| _| || _
| dS ) Initialises the HTMLInputStream.
HTMLInputStream(source, [encoding]) -> Normalized stream from source
for use by html5lib.
source can be either a file-object, local filename or a string.
The optional encoding parameter must be a string that indicates
the encoding. If specified, that encoding will be used,
regardless of any BOM or later declaration (such as in a meta
element)
Nu r r utf-8certain) r supports_lone_surrogatesreportCharacterErrorsr) characterErrorsUCS4characterErrorsUCS2newLineslookupEncoding charEncoding
openStream
dataStreamreset)r& rS r r r r' s
z HTMLUnicodeInputStream.__init__c C s. d| _ d| _d| _g | _d| _d| _d | _d S )N r )r+ chunkSize chunkOffseterrors prevNumLines prevNumCols_bufferedCharacterr7 r r r re s
z HTMLUnicodeInputStream.resetc C s t |dr |}|S t|}|S zvProduces a file object from source.
source can be either a file object, local filename or a string.
r5 )rO r r& rS r# r r r rc s
z!HTMLUnicodeInputStream.openStreamc C sZ | j }|dd|}| j| }|dd|}|dkr#| j| }||fS ||d }||fS )N
r r r )r+ countrj rfindrk )r& r. r+ nLines positionLine lastLinePospositionColumnr r r _position s
z HTMLUnicodeInputStream._positionc C s | | j\}}|d |fS )z:Returns (line, col) of the current position in the stream.r )rv rh )r& linecolr r r r% s z HTMLUnicodeInputStream.positionc C s6 | j | jkr | s tS | j }| j| }|d | _ |S )zo Read one character from the stream or queue if available. Return
EOF when EOF is reached.
r )rh rg readChunkr r+ )r& rh charr r r rz s
zHTMLUnicodeInputStream.charNc C s |d u r| j }| | j\| _| _d| _d| _d| _| j|}| j r,| j | }d | _ n|s0dS t
|dkrWt |d }|dksLd| krJdkrWn n |d | _ |d d }| j r_| | |
d d
}|
d d
}|| _t
|| _d S )
Nrf r Fr r
i z
ro
T)_defaultChunkSizerv rg rj rk r+ rh rd r5 rl r) ordr] replace)r& rg r: lastvr r r ry s0
z HTMLUnicodeInputStream.readChunkc C s( t tt|D ]}| jd q d S )Ninvalid-codepoint)ranger) invalid_unicode_refindallri r9 )r& r: _r r r r^ s z*HTMLUnicodeInputStream.characterErrorsUCS4c C s d}t |D ]Q}|r qt| }| }t|||d r9t|||d }|tv r6| j
d d}q|dkrP|dkrP|t |d krP| j
d qd}| j
d qd S )NF r Tr| i r ) r finditerr groupstartr isSurrogatePairsurrogatePairToCodepointnon_bmp_invalid_codepointsri r9 r) )r& r: skipmatch codepointr* char_valr r r r_ # s" z*HTMLUnicodeInputStream.characterErrorsUCS2Fc
C s zt ||f }W n4 ty< |D ]
}t|dk s J qddd |D }|s-d| }td| }t ||f< Y nw g } || j| j}|du rT| j| j krSn*n|
}|| j krl| | j| j| || _n| | j| jd | s|nq@d|} | S ) z Returns a string of characters from the stream up to but not
including any character in 'characters' or EOF. 'characters' must be
a container that supports the 'in' method and iteration over its
characters.
T rf c S s g | ]}d t | qS )z\x%02x)r )r cr r r r H s z5HTMLUnicodeInputStream.charsUntil..z^%sz[%s]+N)
charsUntilRegExKeyErrorr r< recompiler r+ rh rg endr9 ry )
r&
charactersoppositecharsr regexr> mr rr r r
charsUntil: s:
z!HTMLUnicodeInputStream.charsUntilc C sZ |t ur)| jdkr|| j | _| jd7 _d S | jd8 _| j| j |ks+J d S d S r( )r rh r+ rg )r& rz r r r ungeti s
z HTMLUnicodeInputStream.ungetN)F)rC rD rE rF r~ r' re rc rv r% rz ry r^ r_ r r r r r r rQ s
&
/rQ c @ sR e Zd ZdZ dddZdd Zd d
Zdd d Zd
d Zdd Z dd Z
dS )rR rX N windows-1252Tc C sn | || _t| | j d| _d| _|| _|| _|| _|| _ || _
| || _ | j d dus1J |
dS )rY i d r N)rc rawStreamrQ r' numBytesMetanumBytesChardetoverride_encodingtransport_encodingsame_origin_parent_encodinglikely_encodingdefault_encodingdetermineEncodingrb re )r& rS r r r r r
useChardetr r r r' s z HTMLBinaryInputStream.__init__c C s& | j d j| jd| _t| d S )Nr r )rb
codec_info streamreaderr rd rQ re r7 r r r re s zHTMLBinaryInputStream.resetc C sJ t |dr|}nt|}z
|| W |S ty$ t|}Y |S w rm )rO r r0 r, Exceptionr" rn r r r rc s
z HTMLBinaryInputStream.openStreamc C s | df}|d d ur|S t| jdf}|d d ur |S t| jdf}|d d ur,|S | df}|d d ur:|S t| jdf}|d d urQ|d jdsQ|S t| jdf}|d d ur`|S |rzddl m
} W n t ys Y n@w g }| }|j s| j
| j}t|tsJ |sn
|| || |j r|| t|jd }| j
d |d ur|dfS t| jdf}|d d ur|S tddfS )Nr[ r tentativezutf-16)UniversalDetectorencodingr ) detectBOMra r r detectEncodingMetar name
startswithr %pip._vendor.chardet.universaldetectorr ImportErrordoner r5 r rJ r4 r9 feedcloseresultr0 r )r& chardetrb r buffersdetectorr$ r r r r r sV
z'HTMLBinaryInputStream.determineEncodingc C s | j d dks J t|}|d u rd S |jdv r$td}|d us"J d S || j d kr5| j d df| _ d S | jd |df| _ | td| j d |f )Nr r[ utf-16beutf-16lerZ r z Encoding changed from %s to %s)rb ra r r r0 re r )r& newEncodingr r r changeEncoding s
z$HTMLBinaryInputStream.changeEncodingc
C s t jdt jdt jdt jdt jdi}| jd}t|t s J |
|dd }d}|s?|
|}d}|s?|
|dd }d }|rK| j | t |S | j d
dS ) zAttempts to detect at BOM at the start of the stream. If
an encoding can be determined from the BOM return the name of the
encoding otherwise return NonerZ r r zutf-32lezutf-32be N r r )
codecsBOM_UTF8 BOM_UTF16_LE BOM_UTF16_BE BOM_UTF32_LE BOM_UTF32_BEr r5 rJ r4 getr0 ra )r& bomDictstringr r0 r r r r s&
z HTMLBinaryInputStream.detectBOMc C sV | j | j}t|tsJ t|}| j d | }|dur)|jdv r)t d}|S )z9Report the encoding declared by the meta element
r Nr rZ )
r r5 r rJ r4 EncodingParserr0 getEncodingr ra )r& r$ parserr r r r r 3 s z(HTMLBinaryInputStream.detectEncodingMeta)NNNNr T)T) rC rD rE rF r' re rc r r r r r r r r rR z s
*
> "rR c @ s e Zd ZdZdd Zdd Zdd Zdd Zd
d Zd d
Z dd Z
dd Z e e e
Z
dd Ze eZefddZdd Zdd Zdd Zd S )
EncodingByteszString-like object with an associated position and various extra methods
If the position is ever greater than the string length then an exception is
raisedc C s t |tsJ t| | S r )rJ r4 __new__lowerr& valuer r r r F s zEncodingBytes.__new__c C s
d| _ d S )Nr )rv r r r r r' J s
zEncodingBytes.__init__c C s | S r r r7 r r r __iter__N s zEncodingBytes.__iter__c C s< | j d }| _ |t| krt|dk rt| ||d S )Nr r rv r)
StopIterationrP r& pr r r __next__Q s zEncodingBytes.__next__c C s | S r )r r7 r r r nextY s zEncodingBytes.nextc C s@ | j }|t| kr t|dk rt|d | _ }| ||d S r( r r r r r previous] s zEncodingBytes.previousc C s | j t| kr t|| _ d S r rv r) r )r& r% r r r setPositionf s
zEncodingBytes.setPositionc C s& | j t| kr t| j dkr| j S d S )Nr r r7 r r r getPositionk s
zEncodingBytes.getPositionc C s | | j | j d S Nr )r% r7 r r r getCurrentByteu r8 z EncodingBytes.getCurrentBytec C sR | j }|t| k r$| ||d }||vr|| _|S |d7 }|t| k s || _dS )z Skip past a list of charactersr Nr% r) rv r& r r r r r r r z s zEncodingBytes.skipc C sR | j }|t| k r$| ||d }||v r|| _|S |d7 }|t| k s || _d S r r r r r r skipUntil s zEncodingBytes.skipUntilc C s( | || j}|r| jt|7 _|S )zLook for a sequence of bytes at the start of a string. If the bytes
are found return True and advance the position to the byte after the
match. Otherwise return False and leave the position alone)r r% r) )r& r4 r> r r r
matchBytes s zEncodingBytes.matchBytesc C s6 z| || jt| d | _W dS ty tw )zLook for the next sequence of bytes matching a given sequence. If
a match is found advance the position to the last byte of the matchr T)indexr% r) rv
ValueErrorr r3 r r r jumpTo s zEncodingBytes.jumpToN)rC rD rE rF r r' r r r r r r propertyr% r currentBytespaceCharactersBytesr r r r r r r r r B s"
r c @ sX e Zd ZdZdd Zdd Zdd Zdd Zd
d Zd d
Z dd Z
dd Z dd Z dS )r z?Mini parser for detecting character encoding from meta elementsc C s t || _d| _dS )z3string - the data to work on for encoding detectionN)r r: r r& r: r r r r' s
zEncodingParser.__init__c
C s d| j vrd S d| jfd| jfd| jfd| jfd| jfd| jff}| j D ]?}d}z| j d W n ty< Y | j S w |D ] \}}| j |r\z| }W n
ty[ d}Y nw q?|sc | j S q$| j S ) Ns r: r r7 r r r r s z EncodingParser.handleCommentc C s | j jtvrdS d}d } | }|d u rdS |d dkr/|d dk}|r.|d ur.|| _dS n?|d dkrG|d }t|}|d urF|| _dS n'|d dkrntt|d }| }|d urnt|}|d urn|rl|| _dS |}q
) NTFr s
http-equivr s content-type charsets content) r: r r getAttributer ra ContentAttrParserr parse)r& hasPragmapendingEncodingattrtentativeEncodingcodec
contentParserr r r r s@ zEncodingParser.handleMetac C s
| dS )NF)handlePossibleTagr7 r r r r s
z%EncodingParser.handlePossibleStartTagc C s t | j | dS )NT)r r: r r7 r r r r s
z#EncodingParser.handlePossibleEndTagc C sj | j }|jtvr|r| | dS |t}|dkr#| dS | }|d ur3| }|d us+dS )NTr )r: r asciiLettersBytesr r r spacesAngleBracketsr )r& endTagr: r r r r r r s
z EncodingParser.handlePossibleTagc C r )Nr r r7 r r r r s zEncodingParser.handleOtherc C s | j }|ttdgB }|du st|dksJ |dv r dS g }g } |dkr+|r+n0|tv r4| }n'|dv r?d|dfS |tv rK|| n |du rQdS || t |}q$|dkrj|
d|dfS t | | }|d v r|} t |}||krt | d|d|fS |tv r|| n|| qy|d
krd|dfS |tv r|| n |du rdS || t |}|t v rd|d|fS |tv r|| n |du rdS || q) z_Return a name,value pair for the next attribute in the stream,
if one is found, or None /Nr )r NT =)r
r r; ) ' "r ) r: r r frozensetr) r< asciiUppercaseBytesr9 r r r r )r& r: r attrName attrValue quoteCharr r r r sn
zEncodingParser.getAttributeN)
rC rD rE rF r' r r r r r r r r r r r r r s $ r c @ s e Zd Zdd Zdd ZdS )r c C s t |tsJ || _d S r )rJ r4 r: r r r r r' a s
zContentAttrParser.__init__c C s zy| j d | j jd7 _| j | j jdks W d S | j jd7 _| j | j jdv rS| j j}| j jd7 _| j j}| j |rP| j || j j W S W d S | j j}z| j t | j || j j W W S tyy | j |d Y W S w ty Y d S w )Nr r r )r
r )r: r r% r r r r r )r& quoteMark oldPositionr r r r e s2
zContentAttrParser.parseN)rC rD rE r' r r r r r r ` s r c C s\ t | trz| d} W n
ty Y dS w | dur,zt| W S ty+ Y dS w dS )z{Return the python codec name corresponding to an encoding or None if the
string doesn't correspond to a valid encoding.r N)rJ r4 decodeUnicodeDecodeErrorr
lookupAttributeError)r r r r ra s
ra )0
__future__r r r Zpip._vendor.sixr pip._vendor.six.movesr r r r ior r pip._vendorr
constantsr r
r r r rf r r r r r r invalid_unicode_no_surrogater\ rp r evalr r ascii_punctuation_rer objectr" rW rQ rR r4 r r r ra r r r r sJ
J g Ib = '