Greta 正则表达式Word格式.docx

Greta 正则表达式Word格式.docx

《Greta 正则表达式Word格式.docx》由会员分享，可在线阅读，更多相关《Greta 正则表达式Word格式.docx（34页珍藏版）》请在冰豆网上搜索。

cgroups

match_results,subst_resultsandsplit_results

match_results:

cbackrefs

backref

rstart

rlength

all_backrefs

subst_results:

backref_str

split_results:

strings

TheSyntaxModule

register_intrinsic_charset

CustomizingYourSearch

NOCASE

GLOBAL

MULTILINE

SINGLELINE

EXTENDED

RIGHTMOST

NOBACKREFS

ALLBACKREFS

FIRSTBACKREFS

NORMALIZE

MatchingModes

MODE_FAST

MODE_SAFE

MODE_MIXED

KnownIssuesandPerlIncompatibilities

EmbeddedCodeinaRegularExpression

PatternModifierScope

CommentBlocksBeforeQuantifiers

VariableWidthLook-BehindAssertions

RecursivePatterns

Compile-TimeSwitches

REGEX_WIDE_AND_NARROW

REGEX_POSIX

REGEX_NO_PERL

REGEX_DEBUG

REGEX_DEBUG_HEAP

REGEX_STACK_ALIGNMENT

REGEX_FOLD_INSTANTIATIONS

REGEX_TO_INSTANTIATE

Miscellaneous

StaticConstPatterns

Thread-safety

StackUsage

DBCS

STL

VC7andManagedCode

TemplateInstantiation

ContactInformation

Appendix1:

History

Appendix2:

ImplementationDetails

TheregularexpressiontemplatelibrarycontainsobjectsandfunctionsthatmakeitpossibletoperformpatternmatchingandsubstitutiononstringsinC++. 

Theyare:

∙rpattern:

thepatterntouseduringthesearch.

∙match_results/subst_results:

containerfortheresultsofamatch/substitution.

Toperformasearchorreplaceoperation,youwilltypicallyfirstinitializean 

rpattern 

objectbygivingitastringrepresentingthepatternagainstwhichtomatch. 

Youwillthencallamethodontherpatternobject（match（） 

or 

substitute（）,forinstance）,passingitastringtomatchagainstanda 

match_results 

objectstoreceivetheresultsofthematch. 

Ifthe 

match（）/substitute（） 

fails,themethodreturnsfalse. 

Ifitsucceeds,itreturnstrue,andthe 

objectstorestheresultingarrayof 

backreferences 

internally. 

（Here,theterm 

backreference 

hasthesamemeaningasitdoesinPerl. 

Backreferencesprovideextrainformationaboutwhatpartsofthepatternmatchedwhichpartsofthestring.） 

Therearemethodsonthe 

objecttomakethebackreferenceinformationavailable. 

Forexample:

#include<

iostream>

string>

#include“regexpr2.h”

usingnamespacestd;

usingnamespaceregex;

intmain（）{

match_resultsresults;

stringstr（“Thebookcost$12.34”）;

rpatternpat（“\\$（\\d+）（\\.（\\d\\d））?

”）;

//Matchadollarsignfollowedbyoneormoredigits,

//optionallyfollowedbyaperiodandtwomoredigits.

//Thedouble-escapesarenecessarytosatisfythecompiler.

match_results:

backref_typebr=pat.match（str,results）;

if（br.matched）{

cout<

<

“matchsuccess!

”<

endl;

“price:

”<

br<

}else{

“matchfailed!

}

return0;

}

Theaboveprogramwouldprintoutthefollowing:

matchsuccess!

price:

$12.34

Thefollowingsectionsdiscussthe 

objectindetailandhowtocustomizeyoursearchestobefasterandmoreefficient.

Note:

alldeclarationsintheheaderfile（regexpr2.h）arecontainedinthe 

regex 

namespace. 

Touseanyoftheobjects,methodsorenumerationsdescribedinthisdocument,youmustprependalldeclarationswith“regex:

”oryoumusthavethe“usingnamespaceregex;

”directivesomewherewithintheenclosingscopeofyourdeclarations. 

Forsimplicity,I’veleftoffthe“regex:

”prefixesintherestofmycodesnippets.

Differentregexenginesaregoodondifferenttypesofpatterns. 

Thatsaid,Ihavefoundmyregexenginetobeprettyquick. 

Forabenchmark,Imatchedthepattern“^（[0-9]+）（\-||$）（.*）$”againstthestring“100-thisisalineofftpresponsewhichcontainsamessagestring”. 

GRETAisabout7timesfasterthantheregexlibraryinboost（http:

//www.boost.org）,andabout10timesfasterthantheregularexpressionclassesinATL7. 

Forthisinput,GRETAisevenfasterthanPerl,althoughPerlisfasterforsomeotherpatterns. 

MostregexenginesIhaveseenbuildupanNFA 

（non-deterministicfinitestateautomaton） 

andexecuteititeratively,oftenwithabig,slowswitchstatement. 

Ihaveadifferentapproach:

patternsarecompiledintoadirected,possiblycyclicgraph,andmatchinghappensbytraversingthisgraphrecursively. 

Inaddition,thecodemakesheavyuseoftemplatestofreezethestateoftheflagsintothecompiledpatternsothattheydon’tneedtobecheckedatmatchtime. 

Theresultisaprettyleanblobofcodethatcanmatchyourpatternquickly.

Eventhebestalgorithmshavetheirweaknesses,though. 

MatchingregularexpressionswithbackreferencesisanNP-completeproblem. 

Therearepatternsthatwillmakeanybacktrackingregexenginetakeexponentialtimetofinish. 

（Theseusuallyinvolvenestedquantifiers.） 

Ifyouhaveaperformancecriticalapp,youwouldbesmarttotestyourpatternsforspeed,orprofileyourapptomakesureyouarenotspendingtoomuchtimethrashingaroundintheregexcode. 

You’vebeenwarned!

Also,seethesection 

VC7andManagedCode 

forsomeadviceforcompilingGRETAunderVC7.

Manythingshavechangedsinceversion1.xoftheRegularExpressionTemplateArchive. 

Ifyouhavecodewhichusesversion1.x,youwillnotbeabletouseversion2withoutmakingchangestoyourcode. 

Sorry!

Therewereanumberofunsafe,unintuitiveinterfacefeaturesofversion1thatIfeltwereworthfixingforversion2. 

Ifyouneedversion1,IhaveacopyandI’dbehappytogiveittoyou.

Mostnotably,the 

regexpr 

objecthasgoneaway. 

Itwasasubclassof 

std:

string,with 

match（） 

and 

substitute（） 

methods,anditstoredtheresultsofthematch/substituteinternally. 

Subclassing 

string 

isdangerousbecause 

doesn’thaveavirtualdestructor. 

Also,matchingisconceptuallyaconstoperation,anditseemedwrongthatitshouldchangeinternalstate.

The 

match/count/substitute 

methodshavemovedtothe 

object. 

Thestatethatusedtobestoredinthe 

objectisnowputinamatch_results/subst_results 

container,whichispassedasanoutparametertothe 

match/substitute 

methods.

Also,the 

CSTRINGS 

flaghasgoneaway. 

ItisnolongernecessarytooptimizeapatternforusewithC-styleNULL-terminatedstrings. 

WhenyoupassaC-stylestringtothe 

matchmethod,thesameoptimizationisusedautomatically. 

（Inearly2.Xversionsofthelibrary,therewasa 

basic_rpattern_c 

objectforperformingthisoptimization,butitisnolongernecessaryandhasbeendeprecated.）

Anotherminorchangeinvolvesthe 

register_intrinsic_charset（） 

method. 

Itusedtobeapartof 

rpattern’sinterface,butithasmovedtothesyntaxmodule.

Despitethesweepinginterfacechanges,themajorityoftheback-endcodeisunchanged. 

Youshouldexpectpatternsthatworkedinversion1.xtocontinuetoworkinversion2.

objectcontainstheregularexpressionpatternagainstwhichtomatch. 

Italsoexposesthe 

match（）, 

substitute（）,and 

count（） 

methodsyouwillusetoperformregularexpressionmatches. 

Whenyouinstantiatean 

object,thepatternis“compiled”intoastructurethatspeedsuppatternmatching. 

Oncecompiled,youmayreusethesamepatternformultiplematchoperations.

Hereishow 

isdeclared:

template<

typenameCI,

typenameSY=perl_syntax<

iterator_traits<

CI>

value_type>

>

classbasic_rpattern{

…

};

typedefbasic_rpattern<

basic_string<

TCHAR>

const_iterator>

rpattern;

TCHARconst*>

rpattern_c;

classisatemplateoniteratortype. 

Itisalsoatemplateonthesyntaxmodule. 

Bydefault,thePerlsyntaxmoduleisused,butyouarefreetowriteyourownsyntaxandspecifyitasatemplateparameter. 

Seethesectiononthe 

SyntaxModule.

Thefollowingsectionsdescribethemethodsavailableonthe 

object.

string_type 

isatypedefthatisusedinmanyofthefollowingfunctionprototypes. 

Itisdefinedasfollows:

typedefCIconst_iterator;

typedefstd:

value_typechar_type;

char_type>

string_type;

Thetypedefisalittlecomplicated,butitseffectiswhatyouwouldexpect. 

Iftheresultofdereferencinga 

const_iterator 

isa 

char,then 

isthesameas 

string. 

Ifdereferencinga 

resultsina 

wchar_t,then 

wstring.

Therearetwoconstructorsforinstantiatingan 

Herearetheirprototypes:

rpattern（

const 

&

pat,

REGEX_FLAGSflags=NOFLAGS,

REGEX_MODEmode=MODE_DEFAULT）;

//throw（bad_alloc,bad