SolidSDD - SolidSourceIT
Contents
1. 36 implements NestedThrowabld 37 38 4 The nested 39 protected fin 39 protected final Throwable nested 40 40 4l vs ail y 42 Construct a lt tt gt NestedRuntimeException lt tt gt with t 42 dError lt tt gt with the specifie 43 detail message 43 Annotated source 44 code in file B 45 param msg Detail message age 46 E x 47 public NestedRuntimeException final String msg public NestedError final String msg 48 super msg j 49 this nested null File coverage overview of E clone instance position E Clone 550 Difference 1 6 Selected clone 2 2 One 550 A size content 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 125 When selecting a certain reference file a number of possible cloning partner files are displayed in the second list of the clone selection area see also Figure 2 Using the cloning metrics in these lists one can guide the selection of the most interesting cloning partner highlight C The cloning metrics presented in the second list are relevant for the reference file For example the Cloned column indicates the percentage of the reference file highlight B that is cloned in the partner file highlight C and NOT the other way around Three metrics are used to describe the identifier renaming situation for the selected pair of files ID renaming Indicates whether identifier renaming is exhibited by2. tree action add location Folder X directory action remove location Folder X Folder Y file action add location Folder X Folder Y File2 c node node name Component 27 gt directory action add location Folder X Folder vi file action add location Folder X Folder Z File4 c node lt nodes gt 2007 2010 SolidSource BV www solidsourceit com Solidsource Eod Page 17 This will lead to the following logical hierarchy In detailed steps the above hierarchy is constructed as follows File2 c The whole physical hierarchy is added to Component 1 gt Component 1 Filel c File2 c File3 c File4 c Folder Y is subtracted from Component 1 gt Component 1 File3 c File4 c File2 c is added to Component 1 gt Component 1 File2 c File3 c File4 c Folder Y is added to Component 2 Component 2 Filel c i e File2 c has been previously assigned to Component 1 using a tag with higher precedence file File4 c is added to Compoenent 2 gt Component 2 File c File4 c Component 1 File2 c File3 c i e Component 1 looses File4 c which gets assigned to Component 2 using a tag with higher precedence file 2 4 Loading a project After a duplication detection project has been created and the duplication detection procedure has finished results are saved and they can be inspected at a later time without running the duplication detection again T
3. 2 3 Creating a new project Creating a project is required before running the duplication detection analysis on a new software stack The project enables one to manage the analysis results and to inspect them at a later time without running the duplication detection procedure again Figure 4 New project dialog box New project x gt Project name JBoss M Input Source folder D Source jboss 3 2 8 src Browse o File list optional Browse File filter optional f Discarded folder forced_file java Line filter optional Hierarchy optional Browse M Output Data folder F Analysis boss Browse Analysis settings Language Java Local gap Cumulative gap 2 5 Gap decay 0 1 Minimum clone size 35 TTL Help Create Cancel To create a new duplication detection project choose the File gt New entry from the top menu A dialog box will appear similar to the one depicted in Figure 4 The minimal set of fields that have to be filled in when creating a project is marked by arrows These fields are Project name A user given name to refer to the project in the future Source folder The location of the software stack to be analyzed Data folder The location of the folder where analysis results are to be stored The required amount of space is comparable with the byte size of the software stack to be analyzed When using the Browse b
4. SolidSource Page 16 Table 2 1 Leaf node specification of hierarchy files Tag tree Specifies in a recursive way all files in a given location It takes precedence over previous occurrences of tree That is when a non leaf element uses the tree tag to specify its leaf nodes previous assignments of those leaf nodes to other non leaf nodes done via a tree tag will be discarded directory Specifies all files in a given location but it doesn t work recursively and it takes precedence over tree entries and previous occurrences of directory file Specifies one unique file location and takes precedence over directory and tree entries and previous occurrences of file Action add Indicates that the files referred by the tag should be added to the current selection of the non leaf node remove Indicates that the files referred by the tag should be subtracted from the current selection of the non leaf node Location Gives the physical location of the element referred by the tag enclosed in double quotes The path should be relative to the source folder see Section 2 3 Example Assume the following physical hierarchy with source folder C C Folder_X Folder_Y File1 c C Folder_X Folder_Y File2 c C Folder_X Folder_Z File3 c C Folder_X Folder_Z File4 c And the following hierarchy description lt xml version 1 0 encoding utf 8 gt lt nodes gt lt node name Component 17 gt
5. 8 8 Clone view File view watchlist Blacklist m Fie 11 dusterisre Mo View mode crfaces ClientMethodinterceptorHA java 00 00 Yes 17 common src main org jboss util Abstra Throwable java 3 100 00 Yes 29 common src main org jboss util NestedError java Selected reference File A 3 100 00 Yes 20 common src main org jboss util NestedException java 3 100 00 Yes 31 common src main org jboss util NestedRuntimeException java 3 100 00 Yes 88 j2ee src main javax jms QueueRequestor java 1 1 100 00 Yes 89 j2ee src main iavax ims TopicRequestor java 00 00 hd common src main org jboss util NestedError java common src main org jboss util NestedException java clones Cloned _ 1D Renaming _ Renamings Ren _ Renamed 10s etos 100 00 Yes 10 comp c main org jboss util AbstractNestedThrowable java 4 Cloning metrics file B Jump to next relative to A difference in file A g partner Y ficomman src main org jboss utilNlestedRuntimeException java ficommon src main org jboss utilNlestedError java v Qversion lt tt gt Revision 1 4 2 1 lt tt gt Ni Swap files A and BH y 2 ts FS 3z P ason planet57 com gt Jason pi Gauthor lt a href mailto jason planet57 com Jason Di 33 vs 3 tX 34 public mmm 34 public RE 35 extends 35 extends 36 implements NestedThrowab Jump to next li 3 difference in file B 38 The nested throwable 2 Selected clonin File B
6. 9 Analyzing and managing duplication detection results After the duplication detection process completes see Section 2 5 results are available for inspection A typical analysis scenario involves selecting a set of clone relations and subsequently inspecting the annotated source code and the distribution of the cloning relations see also Section 2 2 2 9 1 Analysis perspectives An analysis can be initiated from three main perspectives e The file perspective e The clone perspective e The system perspective The file perspective In this perspective the File view mode is used for selecting a set of cloning relations follow highlights in Figure 10 First a reference file has to be selected highlight B To this end the user can inspect first the cloning metrics corresponding to each file and included in the list highlight D For a detailed description of these metrics see Section 2 7 Basic list sort operations are available and can be triggered by pressing on the corresponding column headers A user would typically be interested in investigating first clones located in files with a high percentage of cloned statements Figure 10 The SolidCD GUI File view mode lal x Aggregated clone metrics file A _ Clones Fan out Cloned 1D Renaming H SolidCD v1 1 0 trial version F Workspace projects SolidCD SolidCD dist samnles thoss Ihass prj Clone 550 File Tools Settings Help Toolbar eXx oX co
7. SolidSource BV www solidsourceit com SolidSource Page 34 relations will select the corresponding cloning relation in the SolidSDD GUI making it ready for a more detailed inspection Figure 15 Identifying files with a high cloning percentage DEn Fier Status Active ard Types j DEn Dier Statussctive rd Types Fie Coloring atirbuts Cloned 31 x Ordina fiter be Coloring dite Cloned 3 Caption stints name Coloring attrfoute Chorwe X Caption attributs name Colormap att If selecting the relation NotLoggedinException java lt gt UISWTViewEventCancelledExcpetion java in the Azureus example one would obtain a SolidSDD GUI appearance similar to the one depicted in Figure 16 Figure 16 Synchronizing SolidSX and SolidCD Bf solidCD v1 1 0 trial version F Workspace projects SolidCD SolidCD projects azureus azureus prj Clone 424 File Tools Settings Help De MO XIE 6 6 et Clone view File view watchist Blackist 91 38 Yes 1 com aelitis azureus core util bloom impl BloomFilterAddRemovesBit java 1 1 com aelitis azureus core util http HTTPAuthHelper java 2 2 3 81 Partial zl 1 com aelitis azureus core versioncheckiversionCheckClientUDPCodecs java 1 1 41 54 Yes 1 com aelitis azureus core versioncheck VersionCheckClientUDPReply java 1 1 97 87 Yes 1 com aelitis azureus core versioncheckiVersionCheckClientUDPRequest java 1 1 97 87 Yes 1 com aelitis
8. and the type of analysis being performed 1 2 Installation SolidSDD is delivered in binary form with an associated installer that simplifies the installation process When running the installer which is for the largest part self explaining one will be offered the option to select from several installation modes The minimal mode installs only the strictly required components and uses minimal disk space The full mode installs also some examples of datasets from already analyzed projects and can be used to learn the functions of SolidSDD without having to first run the analysis on a specific source tree This mode requires extra disk space to be available 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 7 2 Main functions In this section the main functionality of SolidSDD is described After reading this one should be able to perform a basic scenario create a new project run the duplication detection process and analyze the results To get a better understanding of how SolidSDD can be used to perform real life analysis scenarios the reader is advised to study the use case example presented in Section 3 2 1 Basic code cloning terminology In order to understand how SolidSDD works and how to interpret the results one should be familiar with the duplication detection terminology Here is a minimal set of terms and concepts one should know before using SolidSDD see also Figure 1 Clone An ordered set of state
9. azureus login NotLoggedInException java 2 1 95 24 Yes 1 _ comaelitis azureus plugins extseed ExternalSeedPeer java 6 8 org gudy azureus2 ui swt pluginsimpl UISWTV iewEventCancelledException java 25 26 public class 27 extends Exception 25 26 public class 27 extends Exception 30 d 31 32 ae 33 private static final long serialVersionUID 35 SE 36 37 id 30 public UISWTVieviventCancelledfxcepeion 39 superi TODO Auto generated constructor stub 30 ES 31 32 mp 33 private static final long serialVersionUID 35 ye 36 a 37 Z 30 public Keriegegtupemgeseen 39 super TODO Auto generated constructor stub 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 35 Using the Jump to next difference button see also Figure 10 users can inspect in SolidSDD all differences between the two clone instances of the selected relation In this way they can easily reach the conclusion the two exception classes are almost identical and they only differ by naming i e as opposed to functionality and the hardcoded value of a constant This can be considered a good candidate for refactoring Users can mark it as interesting for further handling by switching to the Clone view mode in SolidSDD and pressing the Move clone to watchlist button of the application toolbar i e button 4 in Figure 3 Subsequentl
10. lines will be treated as non cloned lines so they may generate gaps in the reported cloned code Here is a sample line filter that can be used to discard the first 10 lines of each file during duplication detection def GetSkipLines p_sFile 1_lLines try fileHandle open p sFile r lLines fileHandle readlines fileHandle close except print gt gt sys stderr ERROR could not parse s p_sFile 1_iRange max 10 len lLines 1_1SkipLines range l iRange return 1_1SkipLines 2 3 2 Hierarchy files Hierarchy files are used to give an alternative hierarchical description of the code a logical hierarchy starting from the physical source code tree This can be useful for example when looking at the code from an architectural point of view The logical hierarchy is described using an XML notation The non leaf nodes of the hierarchy have user defined names while the leaf nodes are files in the source tree Non leaf nodes are constructed using node tags lt node name My node gt Children nodes lt node gt The children of non leaf nodes can be other non leaf nodes or leaf nodes Leaf nodes are specified via tree directory or file tags each having two required attributes action and location The definition of each tag and required attribute is given in 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 115 Table 2 1 2007 2010 SolidSource BV www solidsourceit com
11. scrolling as follows scrolling the left panel will automatically scroll the right panel the two panels are linked Scrolling the right panel will not modify the position in the left panel the two panels are not linked The file coverage overview highlight F shows for each file what parts are covered by the selected cloning relations and where identifier renaming took place see also the clone perspective below Besides scrolling one can navigate the annotated code panels by using the Jump to next difference buttons highlight G These will scroll the two panels to the next difference local gap or identifier renaming in the corresponding clone instance of the active cloning relation The clone perspective In this perspective the Clone view mode is used for selecting one cloning relation follow highlights in Figure 12 To this end three lists are available in the clone selection area see also Figure 2 The top list enables one to select a given clone highlight J based on a number of metrics highlight K Besides the clone metrics described in Section 2 7 5 additional metrics are present enabling one to assess the identifier renaming situation for a specific clone These are ID renaming Identifies whether identifier renaming is exhibited by cloning relations between instances of the given clone It can be one of No Yes Partial No indicates no identifier renaming takes place Yes indicates all c
12. the corresponding cloning relation set It can be one of No Yes Partial No indicates no identifier renaming takes place Yes indicates all cloning relations exhibit identifier renaming Partial indicates some relations exhibit identifier renaming but not all of them Renamings Ren The total number of renamings that have been detected Renamed IDs IDs The total number of unique identifiers that have been renamed In general only a few different identifiers are renamed during cloning Large values of this metric can indicate false positives uninteresting clones A default initial selection in the second list is proposed by the application based on the maximum percentage of cloned statements Upon choosing a cloning partner all relevant cloning relations are selected and the annotated code areas are populated see highlight E and Figure 11 A cloning relation is considered as relevant for the chosen pair of files if the corresponding clone instances are located in the chosen files Figure 11 Annotated code areas tomcat src main org jboss web tomcatites WebCtxLoader java gt I fitomestsrc main ora jboss web tomcatites WebCtxLoader java 90 a 74 this ctxLoader new ENCLoader encLoader 91 public void setWarURL URL warURL throws MalformedURL 75 ClassLoader parent encLoader 76 while parent instanceof UnifiedClassLoader f 93 this warURL warURL 77 parent parent getPar
13. two colors are used to mark identifier renaming In the annotated code panel of the reference file renamed identifiers are marked with a green highlight and corresponding identifiers in the partner panel are marked with a yellow highlight highlight S in Figure 12 shows several cases When browsing the code users can jump between locations of renamed identifiers and local gaps using the Jump to next difference buttons located in the upper area of each code panel see highlight G in Figure 10 In the advanced operation mode i e show renaming consistency is checked two additional colors are used dark cyan and red Dark cyan is used in the reference i e left file panel to mark identifiers that are inconsistently renamed That is for one name in the reference panel two or more names are matched in the partner panel after considering all detected renamings In the partner i e right panel identifiers that belong to renamings that brake consistency are marked by red Here is an overview of possible combinations Identifier A Identifier B Situation description reference partner A is consistently renamed to B A is always different from B correspondence can be found in the code below this point Mind the fact that A doesn t have to be different from B or C but B is always different from C A is renamed but not consistently An ID A dark cyan gt ID C red A is renamed but not consistently An ID A d
14. SolidSource se SolidSDD Duplicate Code Detector User Manual For SolidSDD v1 4 Relations Filter selection not listed Coloring attribute Cloned Aggregation type max Direction filter in out Colorize ends start Colormap rainbow Coloring attribute Cloned Caption attribute name Colormap rainbow Legend a Wremotingisrcimainitestidete H Duplication summary Clone 93 tomcat sre main orglibass S wf a jmx src main test compliance a Wemotinglsrcimaintorgljboss Reference testsuite src main org jboss 2 Availabilityltem java Mestsuitelsrcimaintordliboss 2 Duplicate AvgAvailttem java commontsre msinjorajboss utiy commonisrc main org jboss util connector src main org jboss re R oss Renamin gs transaction src main org jboss ti s Fons a lt ttrNestedSQLException lt tt gt with the s Ne message and nested lt tt gt Throwable lt tt gt E AE essage and Mode attons s rar M Detail message param nested Nested lt tt gt Throwable lt tt gt Yr pubic HSE Eine String eg final Tr super msg E this nested nested NestedThrowable Util checkNested this nested Deletions 20 info setProficiency TableColumninfo PROFICIENCY INTERMEDIATE El March 2010 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 13 Contents Nee 9 OL 1 0 EE 5 1 1 S
15. ark cyan gt ID C yellow Dark cyan Red correspondence can be found in the code above this point Mind the fact that A doesn t have to be different from B or C but B is always different from C 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 29 Local gaps are marked in the annotated code areas using a dark grey highlight The system perspective In the system perspective the SolidSX application that is bundled with SolidSDD is used to visualize the overall distribution of the cloning relations and to navigate them follow highlights in Figure 13 For a detailed explanation of the SolidSX GUI and functionality the reader is advised to consult the user manual that comes with the SolidSX distribution If installed the SolidSX application can be invoked from the application toolbar of SolidSDD by pressing the Show clone overview button i e button 7 in Figure 3 If the presence of SolidSX is not automatically detected by SolidSX one can make it available to the application by using the appropriate entry in the Display section of the Settings dialog box see Section 2 6 Figure 13 The SolidSX GUI Ee GET Software Xplorer Standard 1 411 remaining d mi x Ele Edt Tools Yew Heb Node display L Relation display ie F i yg Settings Kaes gd SES J Filter Status Active and Type File TypesFile it il rn Ordinal filfr 45 37
16. ative gap that can be built during comparison see Section 2 1 Allowing the presence of gaps when searching for code clones could lead to identifying relatively very different pieces of code as being similar These pieces could be parts of an original code clone that has been heavily modified or can be simply unrelated code parts yet with a similar construction pattern Such pieces of code are less relevant for immediate consideration and can make the inspection of the overall results very difficult The cumulative gap parameter allows users to control the rate with which local gaps are allowed to become part of code clones during detection This parameter sets a maximum size in SolidSDD statements for the cumulated gap in a clone instance Gap decay The decrease in the computed cumulative gap upon detecting a duplicated statement while comparing clone instances see Section 2 1 Together with the maximum cumulative gap this parameter can be used to control the rate with which gaps are allowed to become part of code clones Minimum clone size The minimum length of the reported clones in SolidSDD statements Clones that are shorter than this will be discarded Setting this threshold value too low can significantly increase the number of reported clones including false positives and makes the result inspection difficult A SolidCD measure of code that tries to abstract from language specific definitions of a statement In general one Soli
17. c Only relations that fall in this interval will be displayed Users can set the limits of the band pass interval by clicking and dragging boundaries in the associated scale visualization highlight F The SolidSDD and SolidSX views are synchronized Clicking on a relation in SolidSX will automatically select that relation in the SolidSDD GUI and the other way around Clicking on hierarchy nodes in the SolidSX visualization will have no effect on the SolidSDD GUI i e no corresponding entries can be identified in this case except when clicking on leaf nodes In that case the corresponding file in the SolidSDD will be selected if it exists i e the file has associated cloning information 2 9 2 Clone management SolidSDD enables users to manage the process of clone inspection via two additional view modes Watchlist and Blacklist see also Section 2 2 These two modes are very similar in layout and functionality with the Clone view mode yet they only work with a subset of the entire collection of detected clones In the Watchlist mode only clones that have been previously marked as interesting are present while in the Blacklist mode only clones marked as not interesting can be inspected The operation of marking clones as interesting or not is further referred to as clone listing 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 31 Clone listing is persistent across analysis sessions and subse
18. cies if changes did take place 2 5 Running the duplication detection To initiate the duplication detection process one can press the Find clones button in the toolbar i e button 1 in Figure 3 or alternatively trigger the process using the Tools gt Find clones menu entry When creating a new project duplication detection is performed automatically The process has two main stages preprocessing and extraction During preprocessing SolidSDD analyzes each file and extracts a number of code features that will be used when searching for clones Depending on the size of the project this stage can be time consuming yet it has limited requirements on the amount of memory available on the processing machine Once retrieved the information is cached The preprocessing stage has two steps the progress of each being displayed by progress bar dialogs e Compiling input list e Pre processing The extraction stage uses the information cached during preprocessing to look for duplicates in the code This stage has 4 steps The progress of each step is displayed using a progress bar dialog The information contained in all files has to be in memory during the extraction step Consequently for large projects this step can be very demanding on the amount of available memory i e RAM To cope with very large projects one could run the duplication detection process on a workstation with sufficient RAM and afterwards distribute the raw
19. dCD statement can be considered to be equivalent with one language specific statement 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 13 In addition to the required fields presented above the New project dialog box contains also three optional fields highlighted by A B C and D in Figure 4 These are not required to perform duplication detection yet they can be used to filter the input or to enable extra analysis features File List The path to a text file containing a list of files to be considered during duplication detection If this parameter is present it will be used to filter source folder The text file should contain the full path of each file to be analyzed on an individual line and should contain files rooted in the source folder File filter A set of wildcard expressions one per line indicating what files from the source folder should be taken into account Each expression should be preceded by a or a indicating whether the files matching the expression should be removed or included in the selection The expression are processed in order from top to bottom and are computed on the list of files in the source folder filtered by the file list filter if present Line filter The filter to be used in order to restrict the duplication detection to a subset of the lines contained in a file This can be used for example to discard automatically generated code from the analysis when
20. de corresponding to the active relation is marked by a light blue background in the annotated code areas Code that belongs to the other selected cloning relations is marked by light green Code in the reference file i e the left panel in the image that is cloned but not covered by any of the selected cloning relations is marked with a yellow background Code that is not cloned can be identified by a white background A cloning relation can be set as active by clicking the left mouse button when the pointer is located in the corresponding green area It is possible that clone instances overlap An indicator situated in the status bar of the application shows the number of overlapping clone instances at the location pointed by the mouse By repeatedly clicking the left mouse button one can cycle through all clone instances and their corresponding cloning relations covering the given location A highlight mechanism is used to indicate the code correspondence in the two panels By hovering with the mouse over a valid statement i e not a comment or white space in the active relation area of one panel the corresponding statement in the other panel will be highlighted It can be that the corresponding statement is situated at another line position making the visual comparison difficult By clicking the left mouse button the two statements are horizontally aligned making the assessment easier The synchronization of the two panels can be affected by
21. e fragments SolidSDD offers an intuitive graphical user interface for assessing the similarity characteristics and the location of the duplicated code fragments in the code stack This enables developers architects to manage the process of refactoring duplicated code by assessing the required effort and establishing priorities The main features for SolidSDD are Fast and scalable SolidSDD can analyze thousands of files in a matter of minutes It uses caches for preprocessed information so the second time it runs even faster In many scenarios it can be executed on the fly enabling quick and timely assessment of code duplication e g before committing code to a versioning repository Tolerant to variations SolidSDD is white space e g comment spacing characters agnostic Additionally it searches beyond exactly duplicated code fragments Most of the times when code is duplicated developers perform small modifications such as renaming variables or inserting deleting some code SolidSDD is able to cope with such modifications while still recognizing the modified piece of code as a duplicate Additionally SolidSDD pinpoints the actual differences in the code so developers can easily assess the required refactoring effort Informative SolidSDD can be used both by developers architects and project managers The detailed code views enable developers to precisely locate the duplicated code fragments the visual overview enable architects t
22. ent 94 String path warURL getFile 78 this delegate UnifiedClassLoader parent 95 File classesDir new File path WEB INF classes 79 36 if classesDir exists 80 81 public void setWarURL URL warURL throws MalformedURL 82 83 this warURL warURL 100 84 String path warURL getFile 101 File libDir new File path WEB INF lib 85 File classesDir new File path WEB INF classes 102 if libDir exists 86 if classesDir exists 103 87 delegate addURL classesDir toURL 104 File jars libDir listFiles 105 int length jars null 2 jars length 0 j lt lend 88 File libDir new File path WEB INF lib if libDir exists File jars libDir listFiles int length jars null jars length 0 for tin zi engthz i for int j 0 36 97 addLifecycleListener LifecycleListener li T 10 Highlight code correspondence ublic LifecycleListener findLifecycleListeners blic LifecycleListene findLifecycleListeners 119 103 120 return new LifecycleListener 0 104 return new LifecycleListener 0 121 105 123 public void removeLifecycleListener LifecycleListener 124 107 public void removeLifecycleListener LifecycleListener Ka 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 26 At any given moment at most one cloning relation out of the selected set is considered to be active The co
23. esults of the cloning detection process such that only relevant cloning relations are taken into account during inspection and reporting see Figure 7 By adjusting the Norm IDs threshold parameter a number of cloning relations can be discarded from analysis and aggregation see Section 2 9 Changes take effect when pressing the Apply button 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 20 Figure 7 The Filters settings Project i Display Report Norm IDs threshold 3 E ee relations Apply Cancel The Display section can be used to adjust the way in which results are presented see Figure 8 The De qualified name length parameter can be used to set the number of path nesting levels presented together with a file name in the GUI For example if De qualified name length 2 files names will include the name of parent folders for the next two levels above in the file hierarchy i e when button 6 in Figure 3 is not pressed The show renaming consistency check box determines the color scheme used to present identifier renaming in the annotated code areas see Section 2 9 Figure 8 The Display settings Location c Program Files SolidSource Solidsx Browse M Other De qualified name length B Show renaming consistency Iv Apply Cancel The Report section can be used to customize the content of generated reports see F
24. ew of reducing the size of the application Two sets of values are given The first set gives the potential benefits of removing all clones The second set gives the total and relative benefit of removing only a given percentage of the top longest clones This percentage is called Easy gain region and can be set on the Report tab of the Project settings dialog see Section 2 6 Physical hierarchy metrics Presents cloning metrics aggregated for the non leaf nodes of the physical hierarchy A detailed description of these metrics is given below Logical hierarchy metrics Presents cloning metrics aggregated for the non leaf nodes of the logical hierarchy if present A detailed description of these metrics is given below File metrics Presents cloning metrics aggregated on file level A detailed description of these metrics is given below 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 22 Clone metrics Presents all detected clones and associated metrics A detailed description of these metrics is given below Physical logical hierarchy metrics aggregate cloning metrics on non leaf nodes of the physical logical hierarchy The metric values of the leaf nodes i e the file level cloning metrics are used as starting point Leaf nodes that have no cloning information available are not taken into account This aspect is particularly relevant when computing average metrics as sums for any given node will be normali
25. f Azureus users would see an image similar to the one depicted in Figure 15 a One component is red util This component has no relations with other components in the system so the duplication causing the high cloning coverage takes place between its sub components By expanding the component i e pressing E while hovering with the mouse over the component or by selecting the component and choosing the Edit gt Expand entry from the top menu users can identify the sub components involved in cloning relations However it may be that from an architectural point of view it is more important to remove cloning relations across components To identify the most promising cloning relations i e corresponding to a high coverage one would have to look for red relations in the image The Ordinal filter field of the relation settings list can facilitate this By selecting the Cloned 96 entry in the Ordinal filter window and dragging the lower border of the scale users can set a filter on the cloning percentage value of the selected relations Only relations with values in the interval set in the window will be displayed By selecting the interval 8096 10096 in the Azureus example two relevant components will be identified azureus and azureus2 see Figure 15 b Right clicking on the relation between them gives a list of all cloning relations that have been detected between files of the two components Clicking on one of the 2007 2010
26. han 3 are worth investigating SolidSDD offers a way to filter out the remaining relations discarding them from the metric aggregation To this end one needs to set the desired threshold for the Norm IDs parameter in the Settings dialog box of the application see Section 2 6 Cloning relations that are included in the aggregation i e Norm IDs lower than the threshold are marked by a blue ball before the name of the file in the cloning partner list highlight P The filtered relations are not marked with a ball and have a faded blue red text color depending on the cause of filtering Red colored relations are automatically filtered via comparison with the threshold Blue colored relations are manually enforced by the user using the pop up menu commands that can be accessed by right clicking with the mouse on a given entry in the partner list The pop up menu also enables users to enforce consideration of interesting relations that have been automatically filtered out In all cases manually enforced filtering takes precedence over the automatically performed one Both in the file and clone perspective identifier renaming is marked using a special color encoding in addition to the one for marking cloned text e depicted in Figure 11 This encoding partially depends on the show renaming consistency parameter setting in the Settings dialog box see Section 2 6 In the basic operation mode i e show renaming consistency is not checked
27. hat pops up An interesting metric for many users is Cloned which gives an overview of the system components containing files with a high percentage of cloned statements Nodes with no information available e g no clones have been detected in the leaf elements are depicted with a grey color To choose a coloring attribute for relations one can click the Coloring attribute entry in the upper right corner highlight and choose one of the available metrics from the list that pops up Interesting metrics in many scenarios are Cloned and Norm IDs see also the file and clone perspectives To filter the displayed relations one can use the Filter and Ordinal filter entries highlight E The Filter entry can be used to visualize either individual cloning relations i e Type Clone or aggregated sets i e Type File This filter can be also used to indicate whether cloning relations filtered based on the Norm IDs value should be included or not in the visualization i e Status Active means that relations that are filtered out will not be displayed One can combine more filters e g Type File and Status Active by holding down the SHIFT key when making the entry selection in the pop up window To filter relations based on a given metric value interval one can use the Ordinal filter highlight F This enables the user to set a band pass interval for a given metri
28. he input and the number of detected clones the generated image can be very crowded To cope with such a situation users can switch to a coarser level of detail Figure 14 Choosing a level of detail for visualizing Azureus clones in SolidSX a The full tree contains many nodes and relations b The collapsed nodes aggregate information and make the image easier to inspect Measured on an Intel Core 2 Duo CPU 2 2 GHz with 4 GB RAM 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 133 To this end one needs to set first the coloring attribute for the nodes of the hierarchy to Level and select a coarser level by clicking with the mouse an entry in the associated scale Levels are numbered using capital letters starting from A for root and finishing with the last letter for the deepest nesting level When selecting a level the corresponding nodes in the visualization are highlighted By pressing the key C or choosing the Edit gt Collapse entry in the top menu of SolidSX the highlighted nodes will be collapsed and will become leaf nodes in the visualization This will reduce the level of detail presented in the hierarchy and will aggregate relations making the image easier to inspect In Figure 14 b level D has been chosen for visualizing clone relations in Azureus The next step of a typical scenario is the relation filtering Not all cloning relations are relevant for a given assessme
29. igure 9 A number of check boxes indicate the sections that will be included in the report The Easy gain region parameter can be used to set a threshold on the amount of clones considered when computing certain project wide metrics see Section 2 7 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 21 Figure 9 The Report settings Project Filters Display iF Included sections IV Project wide metrics Je Physical hierarchy metrics M Logical hierarchy metrics M File metrics M Clone metrics r Other Easy gain region E Yg gi 5 Apply Cancel 2 7 Generating reports SolidSDD can aggregate the cloning information and export it to text reports in the comma separated values CSV format Such reports can be easily imported in spreadsheet processing applications e g MS Office Excel and further embedded in custom reporting systems To generate a report one can press the Generate report button in the toolbar i e button 12 in Figure 3 or alternatively trigger the process using the Tools gt Report menu entry The contents of the generated report can be customized using the Report tab of the Settings dialog box see Section 2 6 Five sections can be included Project wide metrics Presents cloning metrics aggregated for the entire project These metrics try to answer the question What is the potential gain of removing duplicate code from the point of vi
30. instances also referred to as Cloned in the application GUI Identifier renaming Indicates whether the file is involved in cloning relations that exhibit identifier renaming Yes No Partial No indicates no identifier renaming takes place Yes indicates all cloning relations exhibit identifier renaming Partial indicates some relations exhibit identifier renaming but not all of them Clone metrics aggregate raw information resulted during the duplication detection process This is a list of the available clone metrics Clone ID An integer ID for easy reference given to the considered clone by the duplication detection process Clone length The size of ordered set of statements representing the clone i e number of statements in the clone Number of instances Number of clone instances i e the size of the clone set 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 23 Clone fan out Number of files containing instances of the clone Total gap Total number of statements located in gaps in instances of the clone 2 8 Running SolidSDD in batch mode The analysis and reporting features of SolidSDD can be run also in batch mode using the embedded command line functionality To get a listing of the parameters that can be used in batch mode run SolidSDDcmd with the h parameter on the command line 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 124 2
31. loning relations exhibit identifier renaming Partial indicates some relations exhibit identifier renaming but not all of them Renamings Ren The average number of renamings that have been detected for each cloning relation that exhibits renaming i e relations without identifier renaming are 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 127 not taken into account Renamed IDs IDs The average number of unique identifiers that have been renamed for each cloning relation that exhibits renaming i e relations without identifier renaming are not taken into account Norm Ren The average of number of renamings normalized by the length of the corresponding clone instance for each cloning relation that exhibits renaming Norm IDs The average of number of unique identifiers that have been renamed normalized by the length of the corresponding clone instance for each cloning relation that exhibits renaming In general only a few different identifiers are renamed during cloning Large values of this metric i e higher than 3 can indicate false positives uninteresting clones Figure 12 The SolidCD GUI Clone view mode i SolidCD v1 1 0 trial version FWorkspace projectsSolidCD SolidCD distsamplesJbossJboss prj Clone 266 E ni x File Tools Settings Help gt e xX o X t 8 6 68 C Clone view 284 285 286 ee ee Clone listing DEES list
32. lt Cloned lt 100 00 Caption attribute name y Coloring attribute Cloned Colormap rainbow Aggregation type max Direction filter in out Colorize ends start end Colormap green red Relation A EE KE 0 000 coloring metric E Input hierarchy SS source code tree Ordinal edge filter Cloned H SolidSX is a generic visualization tool that visualizes nodes in a hierarchy relations between these nodes and metrics on both The node are visualized as nested boxes placed in concentric rings highlight A whereas relations are depicted as bundled curves highlight B SolidSDD uses SolidSX to visualize physical or logical hierarchies and the associated cloning relations 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 130 SolidSX offers a wide range of selection and filtering scenarios Color plays an important role both nodes and relations can be colored according to associated metrics from the duplication detection process The list of settings available in the upper left corner of SolidSX highlight G can be used to set the coloring attributes for the nodes of the hierarchy and the list in the upper right corner deals with filtering and coloring relations highlight H To choose a coloring attribute for nodes one can click the Coloring attribute entry in the upper left corner highlight D and choose one of the available metrics from the list t
33. ments that is repeated in a number of places in the source code Clone instance A minimal piece of code that includes the ordered set of statements associated with a code clone Clone set The set of all instances of a clone Cloning relation An ordered pair of clone instances belonging to the same clone set The first clone instance of the pair is called the reference the second clone instance is called the cloning partner Clone fan out The number of files containing instances of given clone Local gap A number of neighboring statements in a clone instance that are not part of the ordered statement set of the associated clone Local gaps are the result or code insertion deletion change that typically takes place after duplicating code via copy paste operations Identifier renaming The process of changing the name of identifiers e g variable function type name after duplicating code via copy paste operations Code resulting from copy paste operations is rarely of immediate use during development Such code has to be adapted first to meet the specification that it implements Together with code insertion deletion identifier renaming is a common step performed in combination with code cloning Cumulated gap The sum of all local gaps from the beginning of a clone instance up to a given statement corrected with the gap decay This metric is clone instance and statement specific Gap decay A decrease in the cumulated gap to be conside
34. nt For example when the user is interested in finding files that are linked by cloning relations only one relation between the files is enough to spot them SolidSDD generates aggregate relations for all cloning relations between two files To visualize only these relations one needs to choose the Type File entry in the pop up window displayed by clicking on the Filter field in the upper right corner of the GUI see The system perspective in Section 2 9 1 Additionally one would typically be interested only in relations that have not been filtered out by the automatic comparison with the Norm IDs threshold To remove the filtered out relations from the visualization one needs to add the Status Active entry to the selected filter To do this one should hold down the SHIFT key and select all entries that have to be combined By releasing the SHIFT key the selection will be made and the pop up window discarded Once the appropriate level of detail is chosen and the uninteresting relations removed one can enrich the visualization by adding coloring attributes to nodes and relations In a typical scenario users would be interested in cloning percentage Duplication is easier to remove when it takes place between files with a high cloning percentage To visualize cloning percentage users can choose the Cloned 96 entry in the Coloring attribute field of both node and relation settings lists in the SolidSX GUI see also Figure 13 In the case o
35. nterested and partly instances of not listed clones For files containing only clone instances of clones marked as Green interesting Es For files partly containing instances of clones marked as Light green interesting and partly instances of not listed clones E For files containing instances of both interesting and not Blue interesting clones When the listing mode corresponding to uninteresting clones is not enabled i e button 9 in Figure 3 is not pressed the second list of the clone selection area in the File view mode will not contain entries that would be marked using the first entry in the list above Irrespective of the enabled listing modes the Watchlist mode will list all clones marked as interesting and the Blacklist mode will list all clones marked as not interesting This enables usage scenarios in which an architect inspects first the duplication detection results and marks the clones that need to be removed from the software and subsequently passes the information to a team of developers that refactor the code and remove the corresponding cloning relations To this end both the architect and the developers can use the SolidSDD application and pass refactoring targets via the listing mechanism 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 132 3 Usage example In this section a typical usage example of SolidSDD is presented To illustrate the scenario steps an Open Sou
36. o assess the architectural implications of duplications in order to initiate and prioritize refactoring the aggregated duplication metrics enable managers to take informed decisions regarding maintenance and trigger timely actions Configurable The duplication detection engine of SolidSDD can be configured for various detection strategies and accuracy levels to suit specific analysis contexts and needs Easy to integrate SolidSDD can be integrated in the build environment via its command line interface In this way the duplication detection operation can be executed in batch mode for example together with the nightly builds The recorded end to end analysis time of the Azureus code stack Java application with 3240 files 28MB on a Intel Core 2 Duo CPU 2 2 GHz with 4GB RAM was 68 seconds 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 16 1 1 Supported configurations SolidSDD supports currently the analysis of software stacks written in the C C C or the Java programming languages Other languages can be supported on demand System requirements e Operating system Windows 2000 NT XP or Vista 32 bit e Memory 1GB minimum 4 GB advised D Graphics card OpenGL 1 0 compliant in full color RGBA mode resolution of 1024 x 768 minimum 1280 x 1024 or higher advised H Hard disk space 100 MB free minimum The actual amount of free space required is dependent on the size of the analyzed repository
37. o this end one needs to load the project using the File gt Open menu entry in the top menu When modifying files after running the duplication detection the clone information can be corrupted and become outdated When loading a project SolidSDD initiates a check of consistency i e a progress bar dialog pops up with the caption Checking files This check tries to identify the files that have changed in the meantime and marks the duplication detection results that may have been invalidated in the process Invalid clones and files containing instances of invalid clones are marked via a golden ball in the lists of the clone selection area see Figure 5 This however is no guarantee that the marked information is no longer valid but only a warning To make sure the information is consistent when corruption warnings are issued one needs to run the duplication detection process again Figure 5 Clone consistency warning Clone view File view watchlist Blacklist File containing instances of a corrupted clone 189 ijmxlsrcimainitesthimplementationtutil AgentIDTEST java 190 jmxisrcimain itest performance dynamic InvocationTEST java 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 18 Users can safely skip the consistency step at load time if they are confident files have not been changed in the meantime This will speed up the loading at the risk of introducing presentation inconsisten
38. on used to visualize the cloning relations the appearance of the annotated code and the level of detail used for qualifying file names Report Contains settings for the type and amount of information included in generated reports The Project section is to be used with two main scenarios 1 Importing duplication detection results generated on another machine To this end the directory given by the Data folder parameter see Section 2 3 on the machine used for detection needs to be copied to the target machine Next the actual location of the original input needs to be set i e if this is different by adjusting the Source folder parameter on the target machine Changes take effect when pressing the Apply button 2 Tuning the duplication detection process parameters see Section 2 3 for a detailed description After adjusting these parameters one has to run the duplication detection process again for changes to take effect see Section 2 5 Figure 6 The Project settings Project Fiters Display Report M Input Source folder b Source JBoss Browse File list optional Browse File Filter optional corea core3 disk Line filter optional Hierarchy optional Browse r Detection process Language aa zl Local gap R Cumulative gap es Gap decay fr o Minimum clone size fs Help Apply Cancel The Filters section can be used to filter the r
39. public void Overall clone metrics File view Watchlist Blacklist Yes 16 141 IRE instance 3 39 0 31 metrics 131 D 2 0 04 0 04 Set the ConnectionFactoryInterface value Bparam connectionFactoryInterface Pas Comncc jmx managed attribute s WC 162 Greturn The SARDeployer mbean name 163 jmx managed attribute The ConnectionFactoryImplClass attribute holds the It should be supplied by xslt from ra xml hd gt Clone 266 Clone 266 A When selecting a clone highlight J the application fills the bottom lists of the clone selection area with all instances of the clone set enabling one to specify cloning relations by selecting the clone reference and the partner instances of the pair The left list gives the reference highlight N the right list gives the partner highlight O A number of metrics can be used for making a selection These are similar with the ones descri 2007 2010 SolidSource BV bed above but concern only one cloning relation Modifying the selection of the reference www solidsourceit com SolidSource Page 28 instance affects the contents of the cloning partner list and new metric values will be presented according to the new possible pairs i e cloning relations The Norm IDs identifier can be a very good indicator of false positives uninteresting clone relations Usually relations with a Norm IDs value lower t
40. quent runs of the detection process That is when a given clone is marked as interesting uninteresting during analysis it will keep that classification next time when loading the project and even after running the duplication detection again if the associated clone set does not change Clone listing can be done while in the Clone view mode by using buttons 2 3 4 and 5 of the application toolbar see Figure 3 In this view clones previously marked as interesting will be indicated via a green box preceding the clone ID in the clone selection area while uninteresting clones will be indicated by a red box see highlight L in Figure 12 These boxes and the associated clones are visible in the Clone view mode only when the corresponding listing modes are enabled The listing modes can be enabled disabled using buttons 8 and 9 of the application toolbar see Figure 3 Clones marked as not interesting will be discarded from all metric aggregations e g clone metric aggregation at file level if the corresponding listing mode is not enabled i e button 9 in Figure 3 is not pressed In the File view mode clone listing for the clones with instances in the presented files is suggested via boxes preceding file names and a more complex coloring scheme For files containing only clone instances of clones marked as Red not interesting For files partly containing instances of clones marked as not Light red i
41. rce Software OSS project is used Azureus The source code of Azureus is available at the project web page http azureus sourceforge net This section assumes the reader is familiar with the main concepts way of working and user interface of SolidSDD These notions are detailed in the first two sections of the user manual A typical analysis scenario would start with creating a new project The default parameters can be left unchanged The only mandatory information that has to be provided is the user given project name the location of the source code input the folder where the results will be stored and the programming language of the analyzed code for details see Section 2 3 When a project is created the duplication detection process is run automatically Depending on the size of the input and the chosen project settings the duplication detection process can take from a few seconds to many hours In the case of Azureus the end to end duplication detection process takes approximately 1 minute When the process finished a summary dialog box is displayed reporting the number of clones and the elapsed analysis time After de cloning detection process completes one can start inspecting the results using a system perspective see Section 2 9 1 To this end one has to start the bundled SolidSX application by pressing the Show clone overview button of the SolidSDD application toolbar e button 7 in Figure 3 Depending on the size of t
42. red for each statement of a clone instance that is part of the ordered set of statements of the associated clone This metric is clone instance and location specific being considered on the same set of statements as the cumulated gap 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 8 Figure 1 Basic clone detection terminology if my flag 1 int var1 int MaP2 varl 1 War2 1 Clone int var sum var1 Wef2 int var prod var1 Vef2 T 92uejsu JUO if my_flag 1 Il EE int vari SS int var2 varli 1 18s auo D var2 1 File B Identifier int var sum vari var2 1 2 3 4 5 6 7 renaming 8 9 z 92uejsui euo A yand RE eee See int var prod vari var2 if my flag 1 Dj int vari int E cumulated gap 1 int var sum vari var2_new int var prod vari var2 new cumulated gap 1 5 assume gap decay 0 1 Som Local gap 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 19 2 2 GUI layout SolidSDD comes as a Graphics User Interface GUI with multiple interconnected display areas The main components of the GUI are depicted in Figure 2 These are Top menu A classical window menu that offers access to most of the application highlight A features e g duplication detection project management Toolbar An icon based toolbar see Figure 3 for a magnified view that offers highlight B sho
43. results for further analysis and interpretation to users with less capable workstations The extraction stage uses information cached by the preprocessor when possible Files that have not been changed since the last preprocessing stage are used from the cache when running the duplication detection again Files that changed are sent to the preprocessor first When only a few files change between consecutive runs of the clone detector the preprocessing stage has little influence on the total duration of the duplication detection process 2 6 Adjusting the project settings Once a duplication detection project is created users can adjust many of the initial project and application settings to better suit particular analysis needs at a given moment In order to adjust settings one has to access the Settings dialog box To this end one can press the Settings button in the toolbar i e button 11 in Figure 3 or use the Settings gt Settings menu entry in the top menu The Settings dialog box has four sections see Figure 6 that can be accessed using the corresponding tab in the upper left part of the dialog box Project Contains settings for selecting the input and tuning the duplication detection process 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 119 Filters Contains settings for tuning the filtering of the duplication detection results Display Contains settings for the location of the SolidSX applicati
44. rtcuts to some frequently used operations e g run duplication detection View mode selection A number of tabs for choosing the view mode of the clone selection area highlight C Available view modes are e Clone view three lists enabling the user to select a cloning relation by specifying the clone the clone instance used as reference and the clone instance used as cloning partner e File view two lists enabling the user to select all cloning relations between two files by indicating the file where the reference instances are located and the file where the cloning partner instances are located e Watchlist similar to the Clone view but includes only the clones that have been previously marked as interesting for further inspection e Blacklist similar to the Clone view but includes only the clones that have been previously marked as not interesting for further inspection Clone selection area A number of lists depending on the selected view mode for selecting a highlight D given set of cloning relations for further inspection Reference file area The annotated code of the file containing the reference instance or set of highlight E instances corresponding to the currently selected set of cloning relations Clone partner file area The annotated code of the file containing the cloning partner instance or set highlight F of instances corresponding to the currently selected set of cloning relations In a typical usage
45. scenario the user has to Create a duplication detection project see Section 2 3 or load a previously defined one see Section 2 4 A Adjust the project settings see Section 2 6 if necessary Run tbe duplication detection process see Section 2 5 Select a set of cloning relations in one of the available view modes nspect in detail the cloning metrics and the annotated source code of the reference and cloning partner instances Assess the distribution of the cloning relations see Section 2 9 Generate duplication detection reports see Section 2 7 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 10 A more detailed description of the SolidSDD GUI and clone analysis features is given in Section 2 9 Figure 2 SolidCD GUI layout No cones _ Cloned 1D Renaming Iw No renamings Clone selection reference and cloning partner EH Annotated code o Annotated code of reference instance cloning partner instance Figure 3 Application toolbar Remove clone Remove clone Include watchlist Show duplication Find clones from blacklist from watchlist in the overview summary 1 2 3 45678 10 12 Generate report Move clone Move o Show clone Include blacklist i Settings to blacklist to watchlist filenames overview in the overview 2007 2010 SolidSource BV www solidsourceit com Solidsource s Page 11
46. the position of such code can be programmatically detected The drop list contains all registered filters Filters are scripts written in the Python programming language that use the open SolidSDD API For more information on line filter scripts e g developing and registering line filters see Section 2 3 1 Hierarchy The name of a XML file containing an alternative hierarchical description of the input By default the hierarchy given by the source code tree will be used for aggregating and presenting the analysis results This is called the physical hierarchy In addition to this users can specify a logical decomposition of the system e g from an architectural point of view and use this when analyzing the duplication detection results For more information on the XML schema used to describe the logical hierarchy see Section 2 3 2 During analysis users can switch back and forth between the physical and logical hierarchies 2 3 1 Line filter scripts Line filter scripts are text processing programs written in the Python programming language and implementing a simple line selection interface The only method of this interface has the following signature Name GetSkipLines Input parameters file name name of the file being filtered 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 14 Output parameters A list of all line numbers pointing to lines that will be discarded during duplication detection Discarded
47. upported configurations cccccccececessesssnececeeeceeseenaeseeeeecesseseaaeseceeeceusesaeaeeeseessessesaeaeeeeeeeseeeegs 6 1 2 LUE LE TEE 6 2 Malin PUR ctionis c etre eee ET EE exe Dou eus iae toistaze 7 2 1 Basic code cloning terminology 2 toe ote eate e an ces Eee ORO chics 7 2 2 GUT OU coed cese tees terree i oe d Pepe peccet ee nie Sce LI MM de ei Me Tee Dee 9 2 3 Creating a NeW lge E EE 11 2 3 1 Line diker Serp arrire E E E E EA 13 2 3 2 Hierarchy te geet ete tics e ee EE Deet 14 2 4 Loading ee GE 17 2 5 Running the duplication detection sise 18 2 6 Adj sting ue Ee En d LTC 18 2 7 Generating PED OMS ien eege ees de geed bete eer Est d gedeien bandes de egee E 21 2 8 Running SolidSDD in batch mode ie 23 2 9 Analyzing and managing duplication detection results 24 2 9 1 Analysis perspectives us eene nnne nenne nennnn nnns essent ananas risen seria sans ases nnne nan 24 2 9 2 Clone management EE 30 3 Usag F2 chez UTE 32 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 15 1 Introduction SolidSDD is an application for detecting and managing duplicate code It can be used to analyze large projects and detect code that has been cloned e g via cut n paste operations during development Identifying such code fragments can be very useful for facilitating development and maintenance or for reducing the memory footprint of an application In addition to identifying the duplicated cod
48. utton to select an existing folder form the disk the project name field will be automatically added at the end of the chosen folder name Language The programming language for which the analysis is performed Only the files that match this field will be analyzed File extensions are used to perform the 2007 2010 SolidSource BV www solidsourceit com SolidSource Page 12 matching e g c and h files will be associated with the C programming language In the Analysis settings section the dialog box contains a number of fields for tuning the settings of the duplication detection process These fields are required yet have default values that should lead to satisfactory results in most analysis projects These fields can be adjusted afterwards if needed They are Local gap The maximum size of a local gap in a clone instance see Section 2 1 When cloning code users perform small modifications such as renaming variables or adding deleting small portions of code These modifications lead to differences in the cloned code with respect to the cloning source By accepting gaps when searching for code clones one can minimize the influence of modifications on the duplication detection process and discover clone instances that contain small code modifications This field sets the maximum size in SolidSDD statements of a contiguous modified piece of code that could be accepted as part of a clone Cumulative gap The maximum cumul
49. y the Watchlist mode can be used as TODO list by a team of developers instructed with removing duplicate code 2007 2010 SolidSource BV www solidsourceit com
50. zed with the number of considered leafs instead of the total number of leafs present under the considered node For example for a node with 100 leafs out of which only three have cloning information available the average value of a metric X will be computed as the sum of the X metric values corresponding to the three leafs divided by three Here is a list with the available hierarchy metrics Name The name of the considered non leaf node e g un qualified folder name in the physical hierarchy user defined node name in the logical hierarchy Level The level of the non leaf node in the hierarchy i e the distance from the root Average cloned The average percentage of cloned lines per file Average fan out The average clone fan out per file Total clones The total number of clones affecting files under the non leaf node Average clones Average number of clones per file Total statements The total number of cloned statements in files under the non leaf node Average statements The average number of cloned statements per file File metrics are computed by aggregating the clone information of all clone instances located in a given file Here is a list of the available file metrics File name The qualified name of the considered file Number of clones Number of clones with clone instances in the file Clone fan out Number of files with which the file has cloning relations Clone coverage Percentage of statements located in clone
Download Pdf Manuals
Related Search