原文一：

JavaProgrammingwithOracleJDBC：Performance

Performanceisusuallyconsideredanissueattheendofadevelopmentcyclewhenitshouldreallybeconsideredfromthestart.Often,ataskcalled"performancetuning"isdoneafterthecodingiscomplete,andtheenduserofaprogramcomplainsabouthowlongittakestheprogramtocompleteaparticulartask.Thenetresultofwaitinguntiltheendofthedevelopmentcycletoconsiderperformanceincludestheexpenseoftheadditionaltimerequiredtorecodeaprogramtoimproveitsperformance.It'smyopinionthatperformanceissomethingthatisbestconsideredatthestartofaproject.

WhenitcomestoperformanceissuesconcerningJDBCprogrammingtherearetwomajorfactorstoconsider.ThefirstistheperformanceofthedatabasestructureandtheSQLstatementsusedagainstit.ThesecondistherelativeefficiencyofthedifferentwaysyoucanusetheJDBCinterfacestomanipulateadatabase.

Intermsofthedatabase'sefficiency,youcanusetheEXPLAINPLANfacilitytoexplainhowthedatabase'soptimizerplanstoexecuteyourSQLstatements.Armedwiththisknowledge,youmaydeterminethatadditionalindexesareneeded,orthatyourequireanalternativemeansofselectingthedatayoudesire.

Ontheotherhand,whenitcomestousingJDBC,youneedtoknowaheadoftimetherelativestrengthsandweaknessesofusingauto-commit,SQL92syntax,andaStatementversusaPreparedStatementversusaCallableStatementobject.Inthischapter,we'llexaminetherelativeperformanceofvariousJDBCobjectsusingexampleprogramsthatreporttheamountoftimeittakestoaccomplishagiventask.We'llfirstlookatauto-commit.Next,we'lllookattheimpactoftheSQL92syntaxparser.Thenwe'llstartaseriesofcomparisonsoftheStatementobjectversusthePreparedStatementobjectversustheCallableStatementobject.Atthesametimewe'llalsoexaminetheperformanceoftheOCIversustheThindriverineachsituationtoseeif,asOracle'sclaims,thereisasignificantenoughperformancegainwiththeOCIdriverthatyoushoulduseitinsteadoftheThindriver.Forthemostpart,ourdiscussionswillbebasedontimingdatafor1,000insertsintothetestperformancetableTESTXXXPERF.Thereareseparateprogramsforperformingthese1,000insertsusingtheOCIdriverandtheThindriver.

Theperformancetestprogramsthemselvesareverysimpleandareavailableonlinewiththerestoftheexamplesinthisbook.However,forbrevity,I'llnotshowthecodefortheexamplesinthischapter.I'llonlytalkaboutthem.Althoughtheactualtimingvalueschangefromsystemtosystem,theirrelativevalues,orratiosfromonesystemtoanother,remainconsistent.ThetimingsusedinthischapterweregatheredusingWindows2000.Usingobjectivedatafromtheseprogramsallowsustocometofactualconclusionsonwhichfactorsimproveperformance,ratherthanrelyingonhearsay.

I'msureyou'llbesurprisedattherealityofperformancefortheseobjects,andIhopeyou'llusethisknowledgetoyouradvantage.Let'sgetstartedwithalookatthetestingframeworkusedinthischapter.

ATestingFramework

Forthemostpart,thetestprogramsinthischapterreportthetimingsforinsertingdataintoatable.IpickedanINSERTstatementbecauseiteliminatestheperformancegainofthedatabaseblockbuffersthatmayskewtimingsforanUPDATE,DELETE,orSELECT.

Thetesttableusedintheexampleprogramsinthischapterisasimplerelationaltable.IwantedittohaveaNUMBER,asmallVARCHAR2,alargeVARCHAR2,andaDATEcolumn.TableTESTXXXPERFisdefinedas:

createtableTestXXXPerf(

idnumber,

codevarchar2(30),

descrvarchar2(80),

insert_uservarchar2(30),

insert_datedate)

tablespaceuserspctfree20

storage(initial1Mnext1Mpctincrease0);

altertableTestXXXPerf

addconstraintTestXXXPerf_Pk

primarykey(id)

usingindex

tablespaceuserspctfree20

storage(initial1Mnext1Mpctincrease0);

Theinitialextentsizeusedforthetablemakesitunlikelythatthedatabasewillneedtotakethetimetoallocateanotherextentduringtheexecutionofoneofthetestprograms.Therefore,extentallocationwillnotimpactthetimings.Giventhisbackground,youshouldhaveacontexttounderstandwhatisdoneineachsectionbyeachtestprogram.

Auto-Commit

Bydefault,JDBC'sauto-commitfeatureison,whichmeansthateachSQLstatementiscommittedasitisexecuted.IfmorethanoneSQLstatementisexecutedbyyourprogram,thenasmallperformanceincreasecanbeachievedbyturningoffauto-commit.

Let'stakealookatsomenumbers.

Table19-1

showstheaveragetime,inmilliseconds,neededtoinsert1,000rowsintotheTESTXXXPERFtableusingaStatementobject.Thetimingsrepresenttheaveragefromthreerunsoftheprogram.Bothdriversexperienceapproximatelyaone-secondlossasoverheadforcommittingbetweeneachSQLstatement.Whenyoudividethatonesecondby1,000inserts,youcanseethatturningoffauto-commitsavesapproximately0.001seconds(1millisecond)perSQLstatement.Whilethat'snotinterestingenoughtowritehomeabout,itdoesdemonstratehowauto-commitcanimpactperformance.

Table19-1:Auto-committimings(inmilliseconds)

Auto-commit

OCI

Thin

On

3,712

3,675

Off

2,613

2,594

Clearly,it'smoreimportanttoturnoffauto-commitformanagingmultisteptransactionsthanforgainingperformance.Butonaheavilyloadedsystemwheremanyusersarecommittingtransactions,theamountoftimeittakestoperformcommitscanbecomequitesignificant.Somyrecommendationistoturnoffauto-commitandmanageyourtransactionsmanually.Therestofthetestsinthischapterareperformedwithauto-committurnedoff.

SQL92TokenParsing

Likeauto-commit,SQL92escapesyntaxtokenparsingisonbydefault.Incaseyoudon'trecall,SQL92tokenparsingallowsyoutoembedSQL92escapesyntaxinyourSQLstatements(see"OracleandSQL92EscapeSyntax"inChapter9).Thesestandards-basedsnippetsofsyntaxareparsedbyaJDBCdrivertransformingtheSQLstatementintoitsnativesyntaxforthetargetdatabase.SQL92escapesyntaxallowsyoutomakeyourcodemoreportable--butdoesthisportabilitycomewithacostintermsofperformance?

Table19-2showsthenumberofmillisecondsneededtoinsert1,000rowsintotheTESTXXXPERFtable.TimingsareshownwiththeSQL92escapesyntaxparseronandoffforboththeOCIandThindrivers.Asbefore,thesetimingsrepresenttheresultofthreeprogramrunsaveragedtogether.

Table19-2:SQL92tokenparsertimings(inmilliseconds)

SQL92parser

OCI

Thin

On

2,567

2,514

Off

2,744

2,550

NoticefromTable19-2thatwiththeOCIdriverwelose177millisecondswhenescapesyntaxparsingisturnedoff,andweloseonly37millisecondswhentheparseristurnedoffwiththeThindriver.Theseresultsaretheoppositeofwhatyoumightintuitivelyexpect.ItappearsthatbothdrivershavebeenoptimizedforSQL92parsing,soyoushouldleaveitonforbestperformance.

NowthatyouknowyouneverhavetoworryaboutturningtheSQL92parseroff,let'smoveontosomethingthathassomepotentialforprovidingasubstantialperformanceimprovement.

StatementVersusPreparedStatement

There'sapopularbeliefthatusingaPreparedStatementobjectisfasterthanusingaStatementobject.Afterall,apreparedstatementhastoverifyitsmetadataagainstthedatabaseonlyonce,whileastatementhastodoiteverytime.Sohowcoulditbeanyotherway?Well,thetruthofthematteristhatittakesabout65iterationsofapreparedstatementbeforeitstotaltimeforexecutioncatchesupwithastatement.WhenitcomestowhichSQLstatementobjectperformsbetterundertypicaluse,aStatementoraPreparedStatement,thetruthisthattheStatementobjectyieldsthebestperformance.WhenyouconsiderhowSQLstatementsaretypicallyusedinanapplication--1or2here,maybe10-20(rarelymore)pertransaction--yourealizethataStatementobjectwillperformtheminlesstimethanaPreparedStatementobject.Inthenexttwosections,we'lllookatthisperformanceissuewithrespecttoboththeOCIdriverandtheThindriver.

TheOCIDriver

Table19-3

showsthetimingsinmillisecondsfor1insertand1,000insertsintheTESTXXXPERFtable.TheinsertsaredonefirstusingaStatementobjectandthenaPreparedStatementobject.Ifyoulookattheresultsfor1,000inserts,youmaythinkthatapreparedstatementperformsbetter.Afterall,at1,000inserts,thePreparedStatementobjectisalmosttwiceasfastastheStatementobject,butifyouexamine

Figure19-1

,you'llseeadifferentstory.

Table19-3:OCIdrivertimings(inmilliseconds)

Inserts

Statement

PreparedStatement

1

10

113

1,000

2,804

1,412

Figure19-1isagraphofthetimingsneededtoinsertvaryingnumbersofrowsusingbothaStatementobjectandaPreparedStatementobject.Thenumberofinsertsbeginsat1andclimbsinintervalsof10uptoamaximumof150inserts.Forthisgraphandforthosethatfollow,thelinesthemselvesarepolynomialtrendlineswithafactorof2.Ichosepolynomiallinesinsteadofstraighttrendlinessoyoucanbetterseeachangeintheperformanceasthenumberofinsertsincreases.Ichoseafactorof2sothelineshaveonlyonecurveinthem.Theimportantthingtonoticeaboutthegraphisthatit'snotuntilabout65insertsthatthePreparedStatementobjectoutperformstheStatementobject.65inserts!Clearly,theStatementobjectismoreefficientundertypicalusewhenusingtheOCIdriver.

Figure19-1

TheThinDriver

IfyouexamineTable19-4(whichshowsthesametimingsasforTable19-3,butfortheThindriver)andFigure19-2(whichshowsthedataincrementally),you'llseethattheThindriverfollowsthesamebehaviorastheOCIdriver.However,sincetheStatementobjectstartsoutperformingbetterthanthePreparedStatementobject,ittakesabout125insertsforthePreparedStatementtooutperformStatement.

Table19-4:Thindrivertimings(inmilliseconds)

Inserts

Statement

PreparedStatement

1

10

113

1,000

2,583

1,739

Figure19-2

WhenyouconsidertypicalSQLstatementusage,evenwiththeThindriver,you'llgetbetterperformanceifyouexecuteyourSQLstatementsusingaStatementobjectinsteadofaPreparedStatementobject.Giventhat,youmayask:whyuseaPreparedStatementatall?ItturnsoutthattherearesomereasonswhyyoumightuseaPreparedStatementobjecttoexecuteSQLstatements.First,thereareseveraltypesofoperationsthatyousimplycan'tperformwithoutPreparedStatementobject.Forexample,youmustuseaPreparedStatementobjectifyouwanttouselargeobjectslikeBLOBsorCLOBsorifyouwishtouseobjectSQL.Essentially,youtradesomelossofperformancefortheaddedfunctionalityofusingtheseobjecttechnologies.AsecondreasontouseaPreparedStatementisitssupportforbatching.

Batching

Asyousawintheprevioussection,PreparedStatementobjectseventuallybecomemoreefficientthantheirStatementcounterpartsafter65-125executionsofthesamestatement.Ifyou'regoingtoexecuteagivenSQLstatementalargenumberoftimes,itmakessensefromaperformancestandpointtouseaPreparedStatementobject.Butifyou'rereallygoingtodothatmanyexecutionsofastatement,orperhapsmorethan50,youshouldconsiderbatching.BatchingismoreefficientbecauseitsendsmultipleSQLstatementstotheserveratonetime.AlthoughJDBCdefinesbatchingcapabilityforStatementobjects,OraclesupportsbatchingonlywhenPrepared-Statementobjectsareused.Thismakessomesense.ASQLstatementinaPreparedStatementobjectisparsedonceandcanbereusedmanytimes.Thisnaturallylendsitselftobatching.

TheOCIDriver

Table19-5listsStatementandbatchedPreparedStatementtimings,inmilliseconds,for1insertandfor1,000inserts.Atthelowend,oneinsert,youtakeasmallperformancehitforsupportingbatching.Atthehighend,1,000inserts,you'vegained75%throughput.

Table19-5:OCIdrivertimings(inmilliseconds)

Inserts

Statement

Batched

1

10

117

1,000

2,804

691

IfyouexamineFigure19-3,atrendlineanalysisoftheStatementobjectversusthebatchedPreparedStatementobject,you'llseethatthistime,thebatchedPrepared-StatementobjectbecomesmoreefficientthantheStatementobjectatabout50inserts.Thisisanimprovementoverthepreparedstatementwithoutbatching.

Figure19-3

WARNING:There'sacatchhere.The8.1.6OCIdriverhasadefectbywhichitdoesnotsupportstandardJavabatching,sothenumbersreportedherewerederivedusingOracle'sproprietarybatching.

Now,let'stakealookatbatchinginconjunctionwiththeThindriver.

TheThinDriver

TheThindriverisevenmoreefficientthantheOCIdriverwhenitcomestousingbatchedpreparedstatements.Table19-6showsthetimingsfortheThindriverusingaStatementobjectversusabatchedPreparedStatementobjectinmillisecondsforthespecifiednumberofinserts.

Table19-6:Thindrivertimings(inmilliseconds)

Inserts

Statement

Batched

1

10

117

1,000

2,583

367

TheThindrivertakesthesameperformancehitonthelowend,oneinsert,butgainsawhopping86%improvementonthehighend.Yes,1,000insertsinlessthanasecond!IfyouexamineFigure19-4,you'llseethatwiththeThindriver,theuseofabatchedPreparedStatementobjectbecomesmoreefficientthanaStatementobjectmorequicklythanwiththeOCIdriver--atabout40inserts.

Figure19-4

IfyouintendtoperformmanyiterationsofthesameSQLstatementagainstadatabase,youshouldconsiderbatchingwithaPreparedStatementobject.

We'vefinishedlookingatimprovingtheperformanceofinserts,updates,anddeletes.Nowlet'sseewhatwecandotosqueakoutalittleperformancewhileselectingdata.

PredefinedSELECTStatements

EverytimeyouexecuteaSELECTstatement,theJDBCdrivermakestworoundtripstothedatabase.Onthefirstroundtrip,itretrievesthemetadataforthecolumnsyouareselecting.Onthesecondroundtrip,itretrievestheactualdatayouselected.Withthisinmind,youcanimprovetheperformanceofaSELECTstatementby50%ifyoupredefinetheSelecstatementbyusingOracle'sdefineColumnType()methodwithanOracleStatementobject(see"DefiningColumns"inChapter9).WhenyoupredefineaSELECTstatement,youprovidetheJDBCdriverwiththecolumnmetadatausingthedefineColumnType()method,obviatingtheneedforthedrivertomakearoundtriptothedatabaseforthatinformation.Hence,forasingletonSELECT,youeliminatehalftheworkwhenyoupredefinethestatement.

Table19-7showsthetimingsinmillisecondsrequiredtoselectasinglerowfromtheTESTXXXPERFtable.Timingsareshownforwhenthecolumntypehasbeenpredefinedandwhenithasnotbeenpredefined.TimingsareshownforboththeOCIandThindrivers.AlthoughthedefineColumnType()methodshowslittleimprovementwitheitherdriverinmytest,onaloadednetwork,you'llseeadifferentiationinthetimingsofabout50%.GivenasituationinwhichyouneedtomakeseveraltightcallstothedatabaseusingaStatement,apredefinedSELECTstatementcansaveyouasignificantamountoftime.

Table19-7:Selecttimings(inmilliseconds)

Driver

Statement

defineColumnType()

OCI

13

10

Thin

13

10

Nowthatwe'velookedatauto-commit,SQL92parsing,preparedstatements,andapredefinedSELECT,let'stakealookattheperformanceofcallablestatements.

CallableStatements

Asyoumayrecall,CallableStatementobjectsareusedtoexecutedatabasestoredprocedures.I'vesavedCallableStatementobjectsuntillast,becausetheyaretheslowestperformersofalltheJDBCSQLexecutioninterfaces.Thismaysoundcounterintuitive,becauseit'scommonlybelievedthatcallingstoredproceduresisfasterthanusingSQL,butthat'ssimplynottrue.GivenasimpleSQLstatement,andastoredprocedurecallthataccomplishesthesametask,thesimpleSQLstatementwillalwaysexecutefaster.Why?Becausewiththestoredprocedure,younotonlyhavethetimeneededtoexecutetheSQLstatementbutalsothetimeneededtodealwiththeoverheadoftheprocedurecallitself.

Table19-8liststherelativetime,inmilliseconds,neededtocallthestoredprocedureTESTXXXPERF$.SETTESTXXXPERF（）.ThisstoredprocedureinsertsonerowintothetableTESTXXXPERF.TimingsareprovidedforboththeOCIandThindrivers.Noticethatbothdriversareslowerwheninsertingarowthiswaythanwhenusingeitherastatementorabatchedpreparedstatement(refertoTables19-3through19-6).Commonsensewilltellyouwhy.TheSETTESTXXXPERF（）procedureinsertsarowintothedatabase.ItdoesexactlythesamethingthattheotherJDBCobjectsdidbutwiththeaddedoverheadofaroundtripforexecutingtheremoteprocedurecall.

Table19-8:Storedprocedurecalltimings(inmilliseconds)

Inserts

OCI

Thin

1

113

117

1,000

1,723

1,752

Storedproceduresdohavetheiruses.IfyouhaveacomplextaskthatrequiresseveralSQLstatementstocomplete,andyouencapsulatethoseSQLstatementsintoastoredprocedurethatyouthencallonlyonce,you'llgetbetterperformancethanifyouexecutedeachSQLstatementseparatelyfromyourprogram.Thisperformancegainistheresultofyourprogramnothavingtomovealltherelateddatabackandforthoverthenetwork,whichisoftentheslowestpartofthedatamanipulationprocess.ThisishowstoredproceduresaresupposedtobeusedwithOracle--notasasubstituteforSQL,butasameanstoperformworkwhereitcanbedonemostefficiently.

OCIVersusThinDrivers

Oracle'sdocumentationstatesthatyoushouldusetheOCIdriverformaximumperformanceandtheThindriverformaximumportability.However,IrecommendusingtheThindriverallthetime.Let'stakealookatsomenumbersfromWindows2000.Table19-9listsallthestatisticswe'vecoveredinthischapter.

Table19-9:OCIversusThindrivertimings(inmilliseconds)

Metric

OCI

Thin

1,000insertswithauto-commit

3,712

3,675

1,000insertswithmanualcommit

2,613

2,594

1insertwithStatement

10

10

1,000insertswithStatement

2,804

2,583

1insertwithPreparedStatement

113

113

1,000insertsbatched

1,482

367

SELECT

10

10

PredefinedSELECT

10

10

1insertwithCallableStatement

113

117

1,000insertswithCallableStatement

1,723

1,752

Totals

12,590

11,231

AsyoucanseefromTable19-9,theThindriverclearlyoutperformstheOCIdriverforeverytypeofoperationexceptexecutionsofCallableStatementobjects.OnaUnixplatform,myexperiencehasbeenthattheCallableStatementnumbersaretiltedevenmoreinfavoroftheOCIdriver.Nonetheless,youcanfeelcompletelycomfortableusingtheThindriverinalmostanysetting.TheThindriverhasbeenwell-tunedbyOracle'sJDBCdevelopmentteamtoperformbetterthanitsOCIcounterpart.

原文二：

ExtendingStruts

Introduction

IhaveseenlotofprojectswherethedevelopersimplementedaproprietaryMVCframework,notbecausetheywantedtodosomethingfundamentallydifferentfromStruts,butbecausetheywerenotawareofhowtoextendStruts.YoucangettotalcontrolbydevelopingyourownMVCframework,butitalsomeansyouhavetocommitalotofresourcestoit,somethingthatmaynotbepossibleinprojectswithtightschedules.

Struts

isnotonlyaverypowerfulframework,butalsoveryextensible.YoucanextendStrutsinthreeways.

PlugIn:CreateyourownPlugInclassifyouwanttoexecutesomebusinesslogicatapplicationstartuporshutdown.

RequestProcessor:CreateyourownRequestProcessorifyouwanttoexecutesomebusinesslogicataparticularpointduringtherequest-processingphase.Forexample,youmightextendRequestProcessortocheckthattheuserisloggedinandhehasoneoftherolestoexecuteaparticularactionbeforeexecutingeveryrequest.

ActionServlet:YoucanextendtheActionServletclassifyouwanttoexecuteyourbusinesslogicateitherapplicationstartuporshutdown,orduringrequestprocessing.ButyoushoulduseitonlyincaseswhereneitherPlugInnorRequestProcessorisabletofulfillyourrequirement.

Inthisarticle,wewilluseasampleStrutsapplicationtodemonstratehowtoextendStrutsusingeachofthesethreeapproaches.Downloadablesamplecodeforeachisavailablebelowinthe

Resources

sectionattheendofthisarticle.TwoofthemostsuccessfulexamplesofStrutsextensionsarethe

StrutsValidation

frameworkandthe

Tiles

framework.

IassumethatyouarealreadyfamiliarwiththeStrutsframeworkandknowhowtocreatesimpleapplicationsusingit.Pleaseseethe

Resources

sectionifyouwanttoknowmoreaboutStruts.

PlugIn

AccordingtotheStrutsdocumentation"Apluginisaconfigurationwrapperforamodule-specificresourceorservicethatneedstobenotifiedaboutapplicationstartupandshutdownevents."WhatthismeansisthatyoucancreateaclassimplementingthePlugIninterfacetodosomethingatapplicationstartuporshutdown.

SayIamcreatingawebapplicationwhereIamusingHibernateasthepersistencemechanism,andIwanttoinitializeHibernateassoonastheapplicationstartsup,sothatbythetimemywebapplicationreceivesthefirstrequest,Hibernateisalreadyconfiguredandreadytouse.WealsowanttoclosedownHibernatewhentheapplicationisshuttingdown.WecanimplementthisrequirementwithaHibernatePlugInbyfollowingtwosimplesteps.

CreateaclassimplementingthePlugIninterface,likethis:

publicclassHibernatePlugInimplementsPlugIn{

privateStringconfigFile;

//Thismethodwillbecalledatapplicationshutdowntime

publicvoiddestroy(){

System.out.println("EnteringHibernatePlugIn.destroy()");

//Puthibernatecleanupcodehere

System.out.println("ExitingHibernatePlugIn.destroy()");

}

//Thismethodwillbecalledatapplicationstartuptime

publicvoidinit(ActionServletactionServlet,ModuleConfigconfig)

throwsServletException{

System.out.println("EnteringHibernatePlugIn.init()");

System.out.println("Valueofinitparameter"+

getConfigFile());

System.out.println("ExitingHibernatePlugIn.init()");

}

publicStringgetConfigFile(){

returnname;

}

publicvoidsetConfigFile(Stringstring){

configFile=string;

}

}

TheclassimplementingPlugIninterfacemustimplementtwomethods:init()anddestroy().init()iscalledwhentheapplicationstartsup,anddestroy()iscalledatshutdown.StrutsallowsyoutopassinitparameterstoyourPlugInclass.Forpassingparameters,youhavetocreateJavaBean-typesettermethodsinyourPlugInclassforeveryparameter.InourHibernatePlugInclass,IwantedtopassthenameoftheconfigFileinsteadofhard-codingitintheapplication.

InformStrutsaboutthenewPlugInbyaddingtheselinestostruts-config.xml:

<struts-config>

<!--MessageResources-->

<message-resourcesparameter=

"sample1.resources.ApplicationResources"/>

<plug-inclassName="com.sample.util.HibernatePlugIn">

<set-propertyproperty="configFile"

value="/hibernate.cfg.xml"/>

</plug-in>

</struts-config>

TheclassNameattributeisthefullyqualifiednameoftheclassimplementingthePlugIninterface.Adda<set-property>elementforeveryinitializationparameterwhichyouwanttopasstoyourPlugInclass.Inourexample,Iwantedtopassthenameoftheconfigfile,soIaddedthe<set-property>elementwiththevalueofconfigfilepath.

BoththeTilesandValidatorframeworksusePlugInsforinitializationbyreadingconfigurationfiles.TwomorethingswhichyoucandoinyourPlugInclassare:

Ifyourapplicationdependsonsomeconfigurationfiles,thenyoucanchecktheiravailabilityinthePlugInclassandthrowaServletExceptioniftheconfigurationfileisnotavailable.ThiswillresultinActionServletbecomingunavailable.

ThePlugIninterface'sinit()methodisyourlastchanceifyouwanttochangesomethinginModuleConfig,whichisacollectionofstaticconfigurationinformationthatdescribesaStruts-basedmodule.StrutswillfreezeModuleConfigonceallPlugInsareprocessed.

HowaRequestisProcessed

ActionServletistheonlyservletinStrutsframework,andisresponsibleforhandlingalloftherequests.Wheneveritreceivesarequest,itfirsttriestofindasub-applicationforthecurrentrequest.Onceasub-applicationisfound,itcreatesaRequestProcessorobjectforthatsub-applicationandcallsitsprocess()methodbypassingitHttpServletRequestandHttpServletResponseobjects.

TheRequestPcess()iswheremostoftherequestprocessingtakesplace.Theprocess()methodisimplementedusingtheTemplateMethoddesignpattern,inwhichthereisaseparatemethodforperformingeachstepofrequestprocessing,andallofthosemethodsarecalledinsequencefromtheprocess()method.Forexample,thereareseparatemethodsforfindingtheActionFormclassassociatedwiththecurrentrequest,andcheckingifthecurrentuserhasoneoftherequiredrolestoexecuteactionmapping.Thisgivesustremendousflexibility.TheRequestProcessorclassintheStrutsdistributionprovidesadefaultimplementationforeachoftherequest-processingsteps.Thatmeansyoucanoverrideonlythemethodsthatinterestyou,andusedefaultimplementationsforrestofthemethods.Forexample,bydefaultStrutscallsrequest.isUserInRole()tofindoutiftheuserhasoneoftherolesrequiredtoexecutethecurrentActionMapping,butifyouwanttoqueryadatabaseforthis,thenthenallyouhavetodoisoverridetheprocessRoles()methodandreturntrueorfalse,basedwhethertheuserhastherequiredroleornot.

Firstwewillseehowtheprocess()methodisimplementedbydefault,andthenIwillexplainwhateachmethodinthedefaultRequestProcessorclassdoes,sothatyoucandecidewhatpartsofrequestprocessingyouwanttochange.

publicvoidprocess(HttpServletRequestrequest,

HttpServletResponsere