3 C\o#@sdZddlZddlmZddlmZddlmZmZm Z m Z ddl m Z ddl mZedZejd ejZejd ejZejd Zyed d dWn"ek rejdZdZYnBXddlmZejdjejZdZddlZejd=ddlZe`[ejdZejdZ edZ!edZ"edZ#edZ$edZ%edZ&edZ'edZ(edZ)ed Z*ed!Z+ed"Z,ed#Z-ed$Z.ed%Z/ed&Z0ed'Z1ed(Z2ed)Z3ed*Z4ed+Z5ed,Z6ed-Z7ed.Z8ed/Z9ed0Z:ed1Z;ed2Zed5Z?ed6Z@ed7ZAed8ZBed9ZCed:ZDed;ZEed<ZFed=ZGed>ZHed?ZIed@ZJedAZKedBZLedCZMedDZNedEZOedFZPedGZQe!e9e%e(e1e0e4e:e,e6e-e7e+e5e'e2e)e*e.e/e"e&e#e3e$e8dHZReSdIdJe eRDZTeUeReUeTkstVdKejdLdMjWdNdOeXeRdPdQdRDZYeZeGeIeHe;eLeMeNgZ[eZe;eOeIeNgZ\dSdTZ]dUdVZ^dWdXZ_dYdZZ`d[d\ZaGd]d^d^ebZcGd_d`d`edZeeGdadbdbebZfeGdcddddebZgdedfZhGdgdhdhebZidS)ia jinja2.lexer ~~~~~~~~~~~~ This module implements a Jinja / Python combination lexer. The `Lexer` class provided by this module is used to do some preprocessing for Jinja. On the one hand it filters out invalid operators like the bitshift operators we don't allow in templates. On the other hand it separates template code and python code in expressions. :copyright: (c) 2017 by the Jinja Team. :license: BSD, see LICENSE for more details. N)deque) itemgetter)implements_iteratorintern iteritems text_type)TemplateSyntaxError)LRUCache2z\s+z7('([^'\\]*(?:\\.[^'\\]*)*)'|"([^"\\]*(?:\\.[^"\\]*)*)")z\d+ufööz evalz[a-zA-Z_][a-zA-Z0-9_]*F) _identifierz[\w{0}]+Tzjinja2._identifierz(?z>=srIzoperators droppedz(%s)rAccs|]}tj|VqdS)N)reescape)rExrDrDrH srMcCs t| S)N)len)rLrDrDrHsrO)keycCsL|tkrt|Stdtdtdtdtdtdtdtdt dt d t d t d i j ||S) Nzbegin of commentzend of commentr)zbegin of statement blockzend of statement blockzbegin of print statementzend of print statementzbegin of line statementzend of line statementztemplate data / textzend of template)reverse_operatorsTOKEN_COMMENT_BEGINTOKEN_COMMENT_END TOKEN_COMMENTTOKEN_LINECOMMENTTOKEN_BLOCK_BEGINTOKEN_BLOCK_ENDTOKEN_VARIABLE_BEGINTOKEN_VARIABLE_ENDTOKEN_LINESTATEMENT_BEGINTOKEN_LINESTATEMENT_END TOKEN_DATA TOKEN_EOFget) token_typerDrDrH_describe_token_typesr`cCs|jdkr|jSt|jS)z#Returns a description of the token.r!)typevaluer`)tokenrDrDrHdescribe_tokens rdcCs2d|kr&|jdd\}}|dkr*|Sn|}t|S)z0Like `describe_token` but for token expressions.r@r!)splitr`)exprrarbrDrDrHdescribe_token_exprs rhcCsttj|S)zsCount the number of newline characters in the string. This is useful for extensions that filter a stream. )rN newline_refindall)rbrDrDrHcount_newlinessrkcCstj}t|jd||jft|jd||jft|jd||jfg}|jdk rp|jt|jdd||jf|jdk r|jt|jdd||jfd d t |d d DS) zACompiles all the rules from the environment into a list of rules.r)blockvariableNZ linestatementz ^[ \t\v]*r,z(?:^|(?<=\S))[^\S\r\n]*cSsg|]}|ddqS)reNrD)rErLrDrDrHrIsz!compile_rules..T)reverse) rJrKrNcomment_start_stringblock_start_stringvariable_start_stringline_statement_prefixappendline_comment_prefixsorted) environmenterulesrDrDrH compile_ruless       ryc@s$eZdZdZefddZddZdS)FailurezjClass that raises a `TemplateSyntaxError` if called. Used by the `Lexer` to specify known errors. cCs||_||_dS)N)message error_class)selfr{clsrDrDrH__init__szFailure.__init__cCs|j|j||dS)N)r|r{)r}linenofilenamerDrDrH__call__szFailure.__call__N)__name__ __module__ __qualname____doc__rrrrDrDrDrHrzs rzc@sTeZdZdZfZddedD\ZZZddZ ddZ d d Z d d Z d dZ dS)Tokenz Token class.ccs|]}tt|VqdS)N)propertyr)rErLrDrDrHrMszToken.cCstj||tt||fS)N)tuple__new__rstr)r~rrarbrDrDrHrsz Token.__new__cCs*|jtkrt|jS|jdkr$|jS|jS)Nr!)rarQrb)r}rDrDrH__str__s    z Token.__str__cCs2|j|krdSd|kr.|jdd|j|jgkSdS)zTest a token against a token expression. This can either be a token type or ``'token_type:token_value'``. This can only test against string values and types. Tr@reF)rarfrb)r}rgrDrDrHtests  z Token.testcGs x|D]}|j|rdSqWdS)z(Test against multiple token expressions.TF)r)r}iterablergrDrDrHtest_anys  zToken.test_anycCsd|j|j|jfS)NzToken(%r, %r, %r))rrarb)r}rDrDrH__repr__ szToken.__repr__N)rrrr __slots__rangerrarbrrrrrrDrDrDrHrs rc@s(eZdZdZddZddZddZdS) TokenStreamIteratorz`The iterator for tokenstreams. Iterate over the stream until the eof token is reached. cCs ||_dS)N)stream)r}rrDrDrHrszTokenStreamIterator.__init__cCs|S)NrD)r}rDrDrH__iter__szTokenStreamIterator.__iter__cCs0|jj}|jtkr"|jjtt|j|S)N)rcurrentrar]close StopIterationnext)r}rcrDrDrH__next__s    zTokenStreamIterator.__next__N)rrrrrrrrDrDrDrHrsrc@s~eZdZdZddZddZddZeZedd d d Z d d Z ddZ dddZ ddZ ddZddZddZddZdS) TokenStreamzA token stream is an iterable that yields :class:`Token`\s. The parser however does not iterate over it but calls :meth:`next` to go one token ahead. The current active token is stored as :attr:`current`. cCs>t||_t|_||_||_d|_tdtd|_ t |dS)NFre) iter_iterr_pushedr!rclosedr TOKEN_INITIALrr)r} generatorr!rrDrDrHr/s zTokenStream.__init__cCst|S)N)r)r}rDrDrHr8szTokenStream.__iter__cCst|jp|jjtk S)N)boolrrrar])r}rDrDrH__bool__;szTokenStream.__bool__cCs| S)NrD)rLrDrDrHrO?szTokenStream.z Are we at the end of the stream?)doccCs|jj|dS)z Push a token back to the stream.N)rrs)r}rcrDrDrHpushAszTokenStream.pushcCs"t|}|j}|j|||_|S)zLook at the next token.)rrr)r}Z old_tokenresultrDrDrHlookEs  zTokenStream.lookrecCsxt|D] }t|q WdS)zGot n tokens ahead.N)rr)r}nrLrDrDrHskipMszTokenStream.skipcCs|jj|rt|SdS)zqPerform the token test and return the token if it matched. Otherwise the return value is `None`. N)rrr)r}rgrDrDrHnext_ifRs zTokenStream.next_ifcCs|j|dk S)z8Like :meth:`next_if` but only returns `True` or `False`.N)r)r}rgrDrDrHskip_ifYszTokenStream.skip_ifc CsX|j}|jr|jj|_n:|jjtk rTyt|j|_Wntk rR|jYnX|S)z|Go one token ahead and return the old one. Use the built-in :func:`next` instead of calling this directly. ) rrpopleftrar]rrrr)r}rvrDrDrHr]s zTokenStream.__next__cCs"t|jjtd|_d|_d|_dS)zClose the stream.rNT)rrrr]rr)r}rDrDrHrlszTokenStream.closec Cst|jj|s^t|}|jjtkr:td||jj|j|jtd|t |jf|jj|j|jz|jSt |XdS)z}Expect a given token type and return it. This accepts the same argument as :meth:`jinja2.lexer.Token.test`. z(unexpected end of template, expected %r.zexpected token %r, got %rN) rrrhrar]rrr!rrdr)r}rgrDrDrHexpectrs    zTokenStream.expectN)re)rrrrrrr __nonzero__rZeosrrrrrrrrrDrDrDrHr(s  rc CsZ|j|j|j|j|j|j|j|j|j|j |j |j f }t j |}|dkrVt|}|t |<|S)z(Return a lexer which is probably cached.N)rpblock_end_stringrqvariable_end_stringrocomment_end_stringrrrt trim_blocks lstrip_blocksnewline_sequencekeep_trailing_newline _lexer_cacher^Lexer)rvrPlexerrDrDrH get_lexers" rc@s>eZdZdZddZddZd ddZdd d Zdd d ZdS)ra Class that implements a lexer for a given environment. Automatically created by the environment class, usually you don't have to do that. Note that the lexer is not automatically bound to an environment. Multiple environments can share the same lexer. csdd}tj}ttdfttdfttdftt dft t dft t dfg}t|}|jrTdpVd}i|jr\|d}|d||j}|j|j} || rd|| jdpd7}|j|j} || rd|| jdpd7}|d||j} | j|j} | r d || jdp d} d } d | ||j|||jf} d | ||j| ||jf}| d <|d<nd||j} |j|_|j|_d|ddjd||j| ||j||jfgfdd|Dtdfdf|dtdfgt|d||j||j|fttfdf|dtdfdfgt |d||j||j|ft!dfg|t"|d||j#||j#ft$dfg|t%|d||j| ||j||j|ftt&fdf|dtdfdfgt'|d t(dfg|t)|d!t*t+fdfgi|_,dS)"NcSstj|tjtjBS)N)rJcompileMS)rLrDrDrHrOsz Lexer.__init__..z\n?rr0z^%s(.*)z|%srez(?!%s)z^[ \t]*z%s%s(?!%s)|%s\+?z %s%s%s|%s\+?rlr)z%srootz (.*?)(?:%s)rAz4(?P(?:\s*%s\-|%s)\s*raw\s*(?:\-%s\s*|%s))cs&g|]\}}d||j||fqS)z(?P<%s_begin>\s*%s\-|%s))r^)rErr) prefix_rerDrHrIsz"Lexer.__init__..z#bygroupz.+z(.*?)((?:\-%s\s*|%s)%s)z#popz(.)zMissing end of comment tagz(?:\-%s\s*|%s)%sz \-%s\s*|%sz1(.*?)((?:\s*%s\-|%s)\s*endraw\s*(?:\-%s\s*|%s%s))zMissing end of raw directivez \s*(\n|$)z(.*?)()(?=\n|$))-rJrK whitespace_reTOKEN_WHITESPACEfloat_re TOKEN_FLOAT integer_re TOKEN_INTEGERname_re TOKEN_NAME string_re TOKEN_STRING operator_reTOKEN_OPERATORryrrrpmatchrogrouprqrrjoinrr\rRrrTrSrzrVrWrXrrYTOKEN_RAW_BEGIN TOKEN_RAW_ENDrZr[TOKEN_LINECOMMENT_BEGINrUTOKEN_LINECOMMENT_ENDrx)r}rvcrwZ tag_rulesZroot_tag_rulesZblock_suffix_reZ no_lstrip_reZ block_diffmZ comment_diffZno_variable_reZ lstrip_reZblock_prefix_reZcomment_prefix_rerD)rrHrs         zLexer.__init__cCstj|j|S)z@Called for strings and template data to normalize it to unicode.)rirr)r}rbrDrDrH_normalize_newlines$szLexer._normalize_newlinesNcCs&|j||||}t|j|||||S)zCCalls tokeniter + tokenize and wraps it in a token stream. ) tokeniterrwrap)r}sourcer!rstaterrDrDrHtokenize(szLexer.tokenizec csjxb|D]X\}}}|tkr"qn2|dkr2d}n"|dkrBd}n|dkrPqn|dkrd|j|}n|dkrr|}n|d krt|}tr|j rtd |||n|d kry$|j|d djd djd}WnHtk r}z*t|j ddj }t||||WYdd}~XnXn:|dkr.t |}n&|dkrBt |}n|dkrTt |}t|||VqWdS)zThis is called with the stream as returned by `tokenize` and wraps every token in a :class:`Token` and converts the value. r*r$r+r%r'r(r-keywordr!zInvalid character in identifierr"reasciibackslashreplacezunicode-escaper@Nr rr#)r'r(r)ignored_tokensrr check_ident isidentifierrencodedecode Exceptionrfstripintr operatorsr) r}rr!rrrcrbrwmsgrDrDrHr.sD    "     z Lexer.wrapccsRt|}|j}|jr>|r>x"dD]}|j|r |jdPq Wdj|}d}d}dg} |dk r|dkr|d ksvtd | j|d nd}|j| d!} t|} g} xx| D]l\} }}| j ||}|dkrq| r|d"krqt |t rxt |D]\}}|j tkr|||n|dkrpxt|jD]0\}}|dk r.|||fV||jd7}Pq.Wtd| n8|j|d}|s|tkr|||fV||jd7}qWn|j}|dkrN|dkr| jdnv|dkr| jdn`|dkr| jdnJ|d#krN| s&td||||| j}||krNtd||f||||s^|tkrj|||fV||jd7}|j}|dk r|dkr| jnT|dkrxHt|jD] \}}|dk r| j|PqWtd| n | j||j| d$} n||krtd| |}PqW|| kr0dStd|||f|||qWdS)%zThis method tokenizes the text and returns the tokens in a generator. Use this method if you just want to tokenize a template.    rrrerNrmrlz invalid stateZ_beginr&r%r+z#bygroupz?%r wanted to resolve the token dynamically but no group matchedr#r:r;r8r9r6r7zunexpected '%s'zunexpected '%s', expected '%s'z#popzC%r wanted to resolve the new state dynamically but no group matchedz,%r yielded empty string without stack changezunexpected char %r at %d)rrr)rmrlr)r&r%r+)r;r9r7r)r splitlinesrendswithrsrAssertionErrorrxrNr isinstancer enumerate __class__rzr groupdictcount RuntimeErrorrignore_if_emptyrpopend)r}rr!rrlinesnewlineposrstackZ statetokensZ source_lengthZbalancing_stackregextokensZ new_stateridxrcrPrbr-Z expected_oppos2rDrDrHrWs                                  zLexer.tokeniter)NNN)NN)NN) rrrrrrrrrrDrDrDrHrs  )r)jrrJ collectionsrr#rZjinja2._compatrrrrZjinja2.exceptionsrZ jinja2.utilsr rrUrrrr SyntaxErrorrrZjinja2r formatpatternsysmodulesrriZ TOKEN_ADDZ TOKEN_ASSIGNZ TOKEN_COLONZ TOKEN_COMMAZ TOKEN_DIVZ TOKEN_DOTZTOKEN_EQZTOKEN_FLOORDIVZTOKEN_GTZ TOKEN_GTEQZ TOKEN_LBRACEZTOKEN_LBRACKETZ TOKEN_LPARENZTOKEN_LTZ TOKEN_LTEQZ TOKEN_MODZ TOKEN_MULZTOKEN_NEZ TOKEN_PIPEZ TOKEN_POWZ TOKEN_RBRACEZTOKEN_RBRACKETZ TOKEN_RPARENZTOKEN_SEMICOLONZ TOKEN_SUBZ TOKEN_TILDErrrrrrrVrWrXrYrrrRrSrTrZr[rrrUr\rr]rdictrQrNrrrur frozensetrrr`rdrhrkryobjectrzrrrrrrrDrDrDrHs             +  ^