USING HISTORICAL DATA FROM SOURCE CODE REVISION HI

ABSTRACTTitleofDocument:USINGHISTORICALDATAFROMSOURCECODEREVISIONHISTORIESTODETECTSOURCECODEPROPERTIESChaddCreightonWilliams,DoctorofPhilosophy,2006DirectedBy:ProfessorJeffreyK.Hollingsworth,DepartmentofComputerScienceInthisdissertation,wedescribeseveraltechniquesforusinghistoricaldataminedfromthesourcecoderevisionhistoriesofsoftwareprojectstodetermineimportantpropertiesofthesourcecode.Thesepropertiesarethenusedtoimprovetheresultsofvariousbug-findingtechniquesaswellastoprovidedocumentationtothedeveloper.Wedescribeamethodtominesourcecoderevisionhistories,inthiscaseCVSrepositories,toextractrelevantinformationtobefedintoastaticsourcecodebugfinderforuseinimprovingtheresultsgeneratedbythebugfindingtool.WeapplythistechniquetotheCVSrepositoriesoftwowidelyusedopensourcesoftwareprojects,ApachehttpdandWine.Weshowhowsourcecoderevisionhistorycanbeusedtoreducefalsepositivesfromastaticsourcecodecheckerthatidentifiesthemisuseofvaluesreturnedfromafunctioncall.Amethodofminingsourcecoderevisionhistoriesforthepurposeoflearningaboutprojectspecificidiomsisthendiscussed.Specifically,weshowhowsourcecoderevisionhistorycanbeusedtoidentifypatternsofcallingsequencesthatdescribehowfunctionsinthesoftwareshouldbeusedinrelationtoeachother.Withthisdata,weareabletofindbugsinthesourcecode,documentAPIusageandidentifyrefactoringevents.Inshort,thisdissertationshowsthatitispossibletoautomaticallydeterminemeaningfulpropertiesofthesourcecodefromstudyingsourcecodechangescatalogedinthesoftwarerevisionhistory.USINGHISTORICALDATAFROMSOURCECODEREVISIONHISTORIESTODETECTSOURCECODEPROPERTIESByChaddCreightonWilliamsDissertationsubmittedtotheFacultyoftheGraduateSchooloftheUniversityofMaryland,CollegePark,inpartialfulfillmentoftherequirementsforthedegreeofDoctorofPhilosophy2006AdvisoryCommittee:ProfessorJeffreyK.Hollingsworth,ChairProfessorVictorBasiliAssistantProfessorJeffreyFosterAssistantProfessorMichaelHicksProfessorMartinLoeb©CopyrightbyChaddCreightonWilliams2006iiDedicationFormyparents.iiiAcknowledgementsIfirstneedtothankmyadvisor,Dr.JeffreyK.Hollingsworth.Hisimmensepatienceandsupportmadethisworkpossible.Muchofagraduatestudent’ssuccessisareflectionoftheadvisor,andIhopemyworkreflectswellonDr.Hollingsworth.Myadvisorycommitteealsodeservesaroundofthanks.Theyprovidedinvaluableadviceonhowtoimprovemydissertation.Dr.BillPugh,amemberofmyproposalcommittee,alsoneedstobeincludedinthisgroup.Hisinfluencethroughmyentireresearchcareerwassecondonlytothatofmyadvisor’s.MyfriendshereatMarylandandelsewheredeservethanksforhelpingtokeepmesaneasIwentthroughallthetrialsofgraduateschool.Thesepeopleincludemyhousematesthroughtheyears(allsevenofthem),myfellowDyninstAPIworkersandanyonewhovisitedmyhousesothatIcouldcookchiliorgrillmeat(nothingismoretherapeutic).Theyaretoonumeroustomentionhere.Thereareanumberoffellowgraduatestudentswhodeservespecialmention.MustafaTikir,BrianPostowandByranBuckallgavemewonderfuladvice,basedontheirexperiences,onhowtobeasuccessfulgraduatestudent.WithouttheiradviceI’msureIwouldhaverepeatedalloftheirmistakes,ratherthatjustafewofthem.IhopethatIhavebeenabletooffersimilaradvicetothenewer,youngergraduatestudentsIhavemetoverthelastfewyears.Finally,myfamilyhasbeenextremelysupportivethroughoutmyentireeducationalcareer.Whiletheymayhavebeenmystifiedastowhyittooksolongformetograduate(aconfusionIsharedattimes),myparentsandsisterofferednothingbutencouragementandmuchneededoptimism.ForthisIameternallygrateful.ivTableofContentsDedication.....................................................................................................................iiAcknowledgements......................................................................................................iiiTableofContents.........................................................................................................ivListofTables...............................................................................................................viListofFigures.............................................................................................................viiChapter1:Introduction.................................................................................................1Chapter2:RelatedWork..............................................................................................72.1StaticAnalysis....................................................................................................72.2SoftwareRevisionHistoryMining...................................................................132.3DynamicAnalysis.............................................................................................21Chapter3:Infrastructure.............................................................................................243.1InteractionWithCVS.......................................................................................253.1.1CVSTransactions......................................................................................253.1.2CVSDatainaDatabase...........................