JP5345020B2 - Clone code detection evaluation apparatus, clone code detection evaluation method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently and correctly evaluate and analyze clone codes by selecting clone codes to be evaluated, while preventing excessive detection. <P>SOLUTION: A detection table creation unit 102 registers in a detection table pieces of interval information for predetermined unit data in a file from which clone codes are to be detected, and registers in the detection table pieces of interval information for the predetermined unit data in a file to be compared with the file from which the clone codes are to be detected. If the pieces of interval information between the predetermined unit data in the detection table meet predetermined requirements, a simple clone coupling unit 103 couples together the corresponding pieces of interval information in the detection table for each file. A clone coupling unit 104 extracts the pieces of interval information in the detection table and registers them in a clone code table. A clone code analysis unit 208 evaluates the degree of effect on the quality of files based on the pieces of interval information registered in the clone code table. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a technique applied to a process for detecting a clone code between files and source code evaluation.

  In recent years, a large amount of source code has been described by a large number of people in the development of computer systems. In computer system development projects, projects are also evaluated. In such a development situation, the source code created by one member may be copied (clone) by another member and used.

  The clone code can be generated as follows. First, there is a case where a clone code is generated when an operator copies and pastes the source code of himself or another person. Secondly, there is a case where similar code is automatically generated by an application framework or the like.

  In many cases, the clone code increases the cost of correction at the time of malfunction or function change. Therefore, it is not desirable that there are many clones, and detection of clones is an important index for evaluating a project.

Japanese Patent Laid-Open No. 2003-29978

  However, there is a conventional method for detecting a clone code, but the calculation amount of the detection process is large, and it takes a lot of time to detect the clone code, which is not practical. For example, Patent Document 1 discloses a technique for efficiently detecting a clone code.

  Patent Document 1 discloses an isomorphic pattern detection system that detects isomorphic patterns from text. In the isomorphic pattern detection system disclosed in Patent Document 1, the isomorphism of the target text is detected by lexical analysis, rule transformation, parameter conversion, isomorphism detection, and isomorphism output. However, although the isomorphic pattern detection system disclosed in Patent Document 1 has a high calculation amount of O (N) to O (NlogN), the extracted clone code is not the longest match, so that overdetection increases. There was a problem.

  On the other hand, there are cases where the clone code originally has no influence on the quality and does not need to be evaluated, such as a case caused by the generation of the framework as described above. In this way, there are important clone codes that have an impact on the quality improvement of the source code and less important clone codes that do not affect the quality itself. The target becomes enormous, leading to a reduction in work efficiency.

  Therefore, an object of the present invention is to efficiently and correctly evaluate and analyze clone codes by preventing overdetection and selecting clone codes to be evaluated.

Clones code detection evaluation apparatus of the present invention registers the segment information indicating the segment to the end position from the start position of each data of the first unit in clone code detection target file in the first detection table, the clone First detection table generation for registering section information indicating a section from a start position to an end position for each data of the first unit in a file to be compared with a code detection target file in the first detection table. And section information of the first unit data in the clone code detection target file registered in the first detection table and section information of the first unit data in the comparison target file. when a predetermined condition is satisfied between the front section information of data of the first unit in the first detection table A binding means for coupling to each file, extracts the section information that is coupled to each of the file from the first detection table by said coupling means, said second corresponding to combined section information by the coupling means 1 Based on the clone code table generating means for extracting the clone code which is the second unit data obtained by combining the unit data and registering it in the clone code table, and the clone code registered in the clone code table, And clone code evaluation means for evaluating the importance of the clone code.
The clone code detection / evaluation method of the present invention is a clone code detection / evaluation method by a clone code detection / evaluation apparatus, and shows a section from the start position to the end position for each data of the first unit in a clone code detection target file. registers the section information on the detection table, the clones code detection target file and the detection table section information indicating a section to the end position from the start position of each data of the first unit in the comparison subject to file A detection table generation step to be registered in the first detection table, section information of first unit data in the clone code detection target file registered in the first detection table, and the first unit in the file to be compared when a predetermined condition is satisfied with the section information of the data, the detected tape A coupling step of coupling the section information of the data of the first unit in the Le for each file, extracts the section information that is coupled to each of the files from the detection table by the coupling step, is coupled by said detecting step It was extracted clones encoding the data of the second unit that combines the data of the first unit corresponding to the section information, and clone encoding table generating step of registering the clones code table, registered in the clones code table And a clone code evaluation step for evaluating the importance of the clone code based on the clone code to be executed .
Program of the present invention, clones code detection target section information indicating a section to the end position from the start position of each data of the first unit and registers the detection table in the file, and the clone code detection target file A detection table generation step for registering section information indicating a section from a start position to an end position for each data of the first unit in the file to be compared in the detection table, and registration in the first detection table. When the predetermined condition is satisfied between the section information of the first unit data in the clone code detection target file and the section information of the first unit data in the comparison target file , a coupling step of coupling the section information of the data of the first unit for each file, The second unit data obtained by extracting the section information combined for each file by the combining step from the detection table and combining the data of the first unit corresponding to the section information combined by the combining step. A clone code table generating step for extracting the clone code and registering it in the clone code table; and a clone code evaluation step for evaluating the importance of the clone code based on the clone code registered in the clone code table ; Is executed by a computer.

  In the present invention, clone code can be efficiently evaluated and analyzed efficiently by preventing overdetection and selecting clone codes to be evaluated.

It is a block diagram which shows the functional structure of the clone code detection evaluation apparatus which concerns on the 1st Embodiment of this invention. It is a figure which shows the hardware constitutions of the clone code detection evaluation apparatus which concerns on the 2nd Embodiment of this invention. It is a flowchart which shows the flow of operation | movement of the clone code detection evaluation apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows the flow of operation | movement of the clone code detection evaluation apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows the flow of operation | movement of the clone code detection evaluation apparatus which concerns on the 1st Embodiment of this invention. It is a figure which shows the update condition of a version with the storage condition of each version of each file in a repository. It is a figure which shows the example of a conversion from a source code to a token sequence, and the structural example of a base table. It is a figure for demonstrating the creation process of the detection table by a detection table preparation part. It is a figure for demonstrating a clone simple joining process. It is a figure for demonstrating a clone joint process. It is a figure for demonstrating operation | movement of an inclusion clone deletion part. It is a figure for demonstrating operation | movement of an inclusion clone deletion part. It is a figure for demonstrating a clone simple joining process. It is a figure which shows the structural example of a clone table. It is a block diagram which shows the functional structure of the clone code detection evaluation apparatus which concerns on the 2nd Embodiment of this invention. It is a figure which shows the attribute of the repetition area which occupies the largest area in the clone code detected in 1st Embodiment. It is a figure which shows the evaluation map which shows importance. It is a figure which shows the example of the clone code determined as importance "small".

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments to which the invention is applied will be described in detail with reference to the accompanying drawings.

  FIG. 1 is a block diagram showing a functional configuration of a clone code detection / evaluation apparatus according to the first embodiment of the present invention. As shown in FIG. 1, the clone code detection / evaluation apparatus 100 according to the present embodiment includes a base table creation unit 101, a detection table creation unit 102, a clone simple combination unit 103, a clone combination unit 104, and an included clone deletion unit 105. Is done.

  The base table creation unit 101 converts the source code of the latest file at the time of the last clone code detection into a token string, and converts the token strings converted into n-grams and the interval information of those token strings into the base table. sign up.

  The detection table creation unit 102 converts the source code of the file that is the detection target of the clone code into a token string, and registers the token strings converted into n-grams and the section information of those token strings in the detection table. Then, the detection table creation unit 102 registers the section information registered in the base table in the detection table by a method described later.

  The clone simple combining unit 103 compares the corresponding section information between the records in the detection table, and combines the section information if the compared section information has the same offset between the files.

  The clone combining unit 104 acquires section information for each file from the detection table generated by the clone simple combining unit 103. Then, the clone combining unit 104 overlaps the end of the section information of the preceding token string and the start of the section information of the subsequent token string, or the end information of the preceding token string and the section information of the subsequent token string Is connected to the new detection table, the section information obtained by combining the section information is registered in the new detection table.

  The inclusion clone deletion unit 105 compares the detection table with the new detection table, and when the section information of the detection table is completely included in the section information of the new detection table, that is, the detection table is overdetected. If there is existing section information, the record in which the over-detected section information is registered in the detection table is deleted.

  FIG. 2 is a diagram illustrating a hardware configuration of the clone code detection / evaluation apparatus 100 according to the present embodiment.

  The CPU 201 comprehensively controls each device and controller connected to the system bus. The ROM 203 or the HD (hard disk) 207 has a basic input / output system (BIOS) or an operating system program that is a control program of the CPU 201, and is executed by the clone code detection / evaluation apparatus 100, for example, as shown in FIGS. A processing program and the like are stored.

  In the example of FIG. 2, the HD 207 is configured to be disposed inside the clone code detection / evaluation apparatus 100. However, as another embodiment, a configuration corresponding to the HD 207 is disposed outside the clone code detection / evaluation apparatus 100. It is good also as a structure. In addition, for example, the program for performing the processing illustrated in FIGS. 3-1 to 3-3 according to the present embodiment is recorded on a computer-readable recording medium such as a flexible disk (FD) or a CD-ROM. It is good also as a structure supplied from a recording medium, and good also as a structure supplied via communication media, such as the internet.

  The RAM 202 functions as a main memory, work area, and the like for the CPU 201. The CPU 201 implements various operations by loading a program necessary for execution of processing into the RAM 202 and executing the program.

  The HD 207 and the FD 206 function as an external memory. The CPU 201 implements various operations by loading a program necessary for execution of processing into the RAM 202 and executing the program.

  The disk controller 205 controls access to external memories such as the HD 207 and the FD 206. The communication I / F controller 204 is connected to the Internet or a LAN, and controls communication with the outside by, for example, TCP / IP.

  The display controller 208 controls image display on the display 209.

  The KB controller 210 receives an operation input from the KB (keyboard) 211 and transmits it to the CPU 201. Although not shown, in addition to the KB 211, a pointing device such as a mouse can also be applied to the clone code detection / evaluation apparatus 100 according to the present embodiment as a user operation means.

  The base table creation unit 101, the detection table creation unit 102, the clone simple combination unit 103, the clone combination unit 104, and the included clone deletion unit 105 shown in FIG. 1 are stored in, for example, the HD 207 and loaded into the RAM 202 as necessary. The configuration is realized by a program and a CPU 201 that executes the program.

FIGS. 3A to 3C are flowcharts showing the operation flow of the clone code detection / evaluation apparatus 100 according to the present embodiment. Hereinafter, the operation flow of the clone code detection / evaluation apparatus 100 will be described with reference to the flowcharts shown in FIGS. 3-1 to 3-3. Before the description, the file storage status in the repository will be described with reference to FIG. An example will be described. In the repository, a file that is a target of detection of a clone code by the clone code detection / evaluation apparatus 100 is stored. The repository may be configured in a partial recording area of a recording medium such as an HDD in the clone code detection / evaluation apparatus 100, or may be configured in a DB server or the like connected to the clone code detection / evaluation apparatus 100 via a communication line. May be. Note that the file storage configuration in the repository shown in FIG. 4 is merely an example, and the clone code detection / evaluation apparatus 100 is not limited to the file shown in FIG. Therefore, in the following description, the operation of the clone code detection / evaluation apparatus 100 will be described using the files other than those shown in FIG.
It shall be.

  FIG. 4 is a diagram showing the update status of each version together with the storage status of each version of each file in the repository. In FIG. 4, the passage of time is shown from the top to the bottom of the drawing. In FIG. 4, file 1 (F1) is registered in the repository with version 1 (V1), and there is no subsequent update. File 2 (F2) is registered in the repository with version 1 (V1), and then updated and registered twice with version 2 (V2) and version 3 (V3). That is, as for the file 2, files of versions 1 to 3 are registered. The file 3 (F3) is registered in the repository with version 1, and then updated and registered with version 2, version 3, and version 4. That is, for the file 3, the files of versions 1 to 4 are registered. A broken line 401 represents the time when the clone code detection process was last performed, and a broken line 402 represents the current time. In other words, FIG. 4 shows that the latest file registered in the repository when clone code is detected last is version 1 for file 1, version 2 for file 2, and version 2 for file 3. Yes. FIG. 4 also shows that file 2 has been updated once from version 2 to version 3 and file 3 has been updated twice from version 2 to version 4 since the last clone code was detected. ing.

  Next, the operation of the clone code detection / evaluation apparatus 100 will be described. 3A, the base table creation unit 101 reads the latest file from the repository at the time when the clone code was last detected (step S301). That is, the base table creation unit 101 reads version 1 of file 1, version 2 of file 2, and version 2 of file 3.

  Subsequently, the base table creation unit 101 parses the source code described in each file fetched from the repository, and converts the source code into a token string such as V if void (step S302). The conversion rule here uses ad hoc.

  FIG. 5 is a diagram illustrating a conversion example from a source code to a token string and a configuration example of a base table. As shown in 501 and 502 of FIG. 5, void (source code) is V (token), method1 (source code) is $ (token), int (source code) is I (token), arg (source) Code () is converted to% (token), long (source code) is converted to L (token), I (source code) is converted to% (token), and 10L (source code) is converted to & (token).

  Subsequently, the base table creation unit 101 converts the token string into an n-gram (n = 1, 2, 3,..., N = 10 in FIG. 5), and converts the token string into the n-gram and the file And the token sequence section information in step S303. As indicated by reference numeral 503 in FIG. 5, in the base table, the token string v $ (I%) [L% = is changed from the second column (2, 1) of the second row of the file 1 to the fourth column of the third row ( 3, 4) ((2,1)-(3,4)), token sequence v $ (I%) [L% = is the first column of the 10th row of the file 2 to the ninth column of the 10th row. ((10,1)-(10,9)), token sequence v $ (I%) [L% = is from the first column of the first row to the ninth column of the first row ((1 , 1)-(1, 9)) is registered as section information.

  Also, the token string $ (I%) [L% = & is from the second column of the second row to the fifth column of the third row ((2, 2)-(3, 5)), Section information indicates that $ (I%) [L% = & exists from the second column of the first row to the tenth column of the first row ((1,2)-(1,10)). It is registered.

  Further, the token string (I%) [L% = &; exists from the third column of the first row of the file 3 to the eleventh column of the first row ((1,3)-(1,11)). Is registered as section information. Further, the token string = $ (%,%,%); exists from the first column of the 12th row to the 12th column of the 12th row ((12,1)-(12,12)). Is registered as section information.

  FIG. 6 is a diagram for explaining detection table creation processing by the detection table creation unit 102. FIG. 6 shows a configuration example of a detection table created when version 3 of the file 3 is a clone code detection target.

  Subsequently, the detection table creation unit 102 determines whether or not an old version of the file 3 to be detected exists in the base table (step S304). When the old version of the file 3 exists, the detection table creation unit 102 deletes the section information of the old version of the detection target file 3 from the base table as indicated by 601 in FIG. 6 (step S305). If this is a file of the same type with a different version, it is not appropriate to detect the common part as a clone code because the new version file is partially updated from the old version file and there are many common parts. Because. On the other hand, if the old version of the file 3 does not exist, the process proceeds to step S306 without executing step S305.

  Subsequently, the detection table creation unit 102 reads version 3 of the file 3 from the repository (step S306). Then, the detection table creation unit 102 parses the source code described in the version 3 of the file 3 and converts it into a token string (step S307).

  Subsequently, the detection table creation unit 102 converts the token string into an n-gram, and registers the token string converted into an n-gram and section information in the file of the corresponding token string in the detection table (step S308). As indicated by 602 in FIG. 6, the clone column v $ (I%) [L% = is from the fifth column of the second row of the version v3 of the file F3 to the thirteenth row of the second row ((2, 5) -(2,13)), clone string $ (I%) [L% = & is the second line of file F3 version v3 from the 6th line to the 14th line ((2,6)- (2,14)), the clone column (I%) [L% = &; is from the seventh row of the second row to the fifteenth column of the second row of version v3 of the file F3 ((2,7)-( 2, 15)) Existence is registered as section information.

  Subsequently, when the section information is registered in the base table in the same token string as the token string in which the section information is registered in the detection table, the detection table creation unit 102 displays the section information of the corresponding token string in the base table. Register in the detection table (step S309).

  As indicated by reference numeral 603 in FIG. 6, in the detection table, three token sequences v $ (I%) [L% =, $ (I%) [L% = &, (I%) [L% = &; Section information is registered, but in the base table, section information of two token sequences v $ (I%) [L% =, $ (I%) [L% = & is registered. ing. Therefore, the detection table creation unit 102 registers the section information of the two token strings v $ (I%) [L% =, $ (I%) [L% = &] in the base table in the detection table.

  Subsequently, in FIG. 3B, the detection table creation unit 102 determines whether or not there is a token string record in which only one section information is registered between files (step S310). When only one section information is registered between files, the detection table creation unit 102 deletes the corresponding record (step S311). This is because it is clear that a token string in which only one piece of section information is registered between files is not a clone code at that time. In the example of FIG. 6, only the section information of the file 3 is registered in the record of the token string (I%) [L% = &;. Therefore, the record of the token string (I%) [L% = &; is deleted. As described above, the detection table indicated by reference numeral 603 in FIG. 6 is created. On the other hand, if there is no token sequence record in which only one section information is registered between files, the process proceeds to step S312 without executing step S311.

  Subsequently, the clone simple combining unit 103 classifies each record of the detection table according to the number of section information (step S312). In the example of FIG. 7, the classification of the record of token sequence V $ (I%) [L% = and the record of token sequence (I%) [L% = &; The token sequence $ (I%) [L% = & record and the token sequence I%) [L% = &;% record in which three pieces of section information are registered are classified.

  Subsequently, the clone simple combining unit 103 performs a clone simple combining process (step S313). That is, the simple clone combining unit 103 compares corresponding section information between records for each classification, and combines the section information if the compared section information has the same offset between files.

  FIG. 7 is a diagram for explaining the process of simple clone combining by the clone simple combining unit 103. Here, as shown by reference numeral 701 in FIG. 7, an example is shown in which section information is registered for four token strings in the detection table. That is, in the detection table, the section information of the token string V $ (I%) [L% = is registered in the four files 4 to 7, and the token string (I%) [L% = & is stored in the four files 4 to 7. ; Section information is registered. In addition, the section information of token sequence $ (I%) [L% = & is registered in the three files 4 to 6, and the token sequence I%) [L% = &;% section information is stored in the three files 4 to 6. Is registered.

  In the example of FIG. 7, as the section information of the token string V $ (I%) [L% =, (2, 1)-(3, 3), ( 3,2)-(4,4), (4,5)-(5,7), (6,2)-(7,4) are registered, and the token string (I%) [L% = &; (2, 2)-(3,4), (3, 3)-(4, 5), (4, 6)-(5, 8), (6, 3)-(7, 5) is registered. All section information to be compared has the same offset (one coordinate in the horizontal axis direction). Therefore, in this case, the section information is combined. In other words, the detection table creation unit 102 generates section information in which the beginning token end of the preceding token string to the end of the latter token string of the compared token strings are combined. At this time, the clone simple combining unit 103 updates the token string of one record among the records divided for each classification to the combined token string, and converts each section information of the record to the combined section information. Update. Then, the clone simple combination unit 103 deletes all other records belonging to the classification. As described above, after the clone simple combining process, a detection table indicated by 702 in FIG. 7 is generated.

  When the section information is combined for each file, the token string corresponding to the combined section information can be regarded as the same token string across a plurality of files. Therefore, in the above-described clone simple combination process, the same number of section information of a certain token string is registered across a plurality of files between records, and the section information of each token string is recorded between records of the same classification. This is performed when a plurality of files are registered with the same offset.

  Subsequently, the clone combining unit 104 performs a clone combining process (step S314). That is, the clone combining unit 104 acquires section information for each file from the detection table generated by the clone simple combining unit 103. Then, the clone combining unit 104 overlaps the end of the section information of the preceding token string and the start of the section information of the subsequent token string, or the end information of the preceding token string and the section information of the subsequent token string Is connected to the new detection table, the section information obtained by combining the section information is registered in the new detection table. As a result, a new detection table is created in which section information corresponding to a series of token strings obtained by combining token strings converted into n-grams is created.

  FIG. 8 is a diagram for explaining clone joining processing by the clone joining unit 104. Reference numeral 801 in FIG. 8 is a detection table, and reference numeral 802 in FIG. 8 is a new detection table generated by performing clone combination processing on the detection table. In the detection table, the section information of the file 9 of the token sequence $ (); (L% = N; is (10, 3)-(11, 5), and the token sequence $ (); [L% = $ ( The section information of the file 9 is (9,5)-(10,7) Therefore, the end of the section information ((9,5)-(10,7)) of the preceding token string and the token information of the latter stage Section information ((10,3)-(11,5)) overlaps with the beginning of the section information, and section information (9,5)-(11,5) obtained by combining these section information is calculated.

  Further, since the section information of the file 9 of the token sequence $ (%,%,%,%) is (11, 5)-(12, 7), the file of the token sequence $ (); (L% = N; 9 is connected to the end of the section information ((10, 3)-(11, 5)), so section information (10, 3)-(12, 7) is calculated by combining the section information. The section information is similarly combined for the other files 10 to 12. Since section information is registered only in the token string $ (); [L% = N; Are not combined.

  Subsequently, the clone combining unit 104 determines whether there is section information combined by the clone combining process (step S315). If there is section information combined by the clone combining process, the clone combining unit 104 registers the combined section information in association with the combined token string in the new detection table (step S316). On the other hand, when there is no section information joined by the clone joining process, the process proceeds to step S317 in FIG. 3-3 without executing step S316.

  Subsequently, the clone combining unit 104 determines whether there is a record that has only one piece of section information registered between files in the new detection table (step S317). If there is a record that has only one section information registered between files in the new detection table, the clone combining unit 104 deletes the corresponding record (step S318). This is because a token string in which only one section information is registered between files is not a clone. In the example of FIG. 8, the record of token sequence $ (); [L% = $ (); [L% = $ (will be deleted. If there is no record that can only be registered, the process proceeds to step S319 without executing step S318, and the process in steps S314 to S318 described above is performed to detect the new table 802 in FIG. A detection table is generated.

  9 and 10 are diagrams for explaining the operation of the inclusion clone deletion unit 105. FIG. The included clone deletion unit 105 separates the files registered in the detection table into groups, and separates the files registered in the new detection table into groups (step S319). In this grouping, files in which section information is registered with the same token string are grouped into the same group. Grouping in this way can reduce the amount of processing by performing overdetection processing for each group having a higher possibility of clones than performing clone overdetection processing described later for each file. Because. 901 in FIG. 9 indicates the status of grouping in the detection table, and 902 in FIG. 9 indicates the status of grouping in the new detection table. Both the detection table and the new detection table are divided into group A of file 9 and file 10 and group B of file 11 and file 12.

  Subsequently, the inclusion clone deletion unit 105 compares the section information for each group between the detection table and the new detection table, and whether or not the section information of the combined table is completely included in the section information of the new combined table. Is determined (step S320). That is, in step S320, it is determined whether or not there is section information that is over-detected in the detection table. For example, when the token string ABABA is registered in the new detection table and the token string BAB is registered in the detection table with section information included in the section information of the token string ABABA, only the longest token string ABABA is stored. The token string BAB included in the token string need not be detected. In such a case, detecting the token string BAB is overdetection, and in step S320, this overdetection is determined.

  If the section information of the combined table is completely included in the section information of the new combined table, the included clone deletion unit 105 deletes the corresponding record from the combined table (step S321). That is, the included clone deletion unit 105 deletes a record in which overdetected section information is registered in the combined table.

  As indicated by 1001 in FIG. 10, the section information of the file 9 of the token sequence $ (%,%,%,%) of the join table is (11,5)-(12,7), and the token sequence $ of the join table is also the same. The section information of the file 10 of (%,%,%,%) is (12, 5)-(13, 7). Further, as indicated by 1002 in FIG. 10, the section information of the file 9 in the token string $ (); [L% = N; $ (%,%,%,%) of the new join table is (10, 3) − (12, 7), similarly, the section information of the file 10 in the token sequence $ (); [L% = N; $ (%,%,%,%) of the new join table is (11, 3)-(13, 7 ). In this way, the interval information of token $ (%,%,%,%) in the join table is the interval of token $ (); [L% = N; $ (%,%,%,%) in the new join table Since it is completely included in the information, the record of token $ (%,%,%,%) in the join table is deleted. On the other hand, when the section information of the join table is not completely included in the section information of the new join table, the process proceeds to step S322 without executing step S321.

  Subsequently, the included clone deletion unit 105 extracts all the records in the detection table and generates a clone table (step S322). Therefore, as indicated by 1001 in FIG. 10, the clone table is composed of records of token sequence $ (); [L% = N; and records of token sequence $ (); [L% = $ ( It will be.

  Subsequently, the included clone deletion unit 105 determines whether or not a new detection table has been generated (step S323). In the example of FIG. 10, since a new detection table is generated, the process proceeds to step S324 in FIG. The clone simple combining unit 103 replaces the new detection table with the detection table (step S324). On the other hand, if a new detection table has not been generated, the process ends.

  The inclusion clone deletion unit 105 executes the clone simple combination process of step S313 on the detection table replaced from the new detection table by the inclusion clone deletion unit 105. That is, the detection table replaced from the new detection table 1101 in FIG. 11 is the section information of token strings $ (); [L% = $ (); [L% = N; (9,5)-(11,5), (10,5)-(12,5) are registered as token strings $ (); [L% = N; $ (%,%,%,%) (10,3)-(12,7), (11,3)-(13,7) are registered as the section information. Therefore, all the section information to be compared has the same offset (one coordinate in the vertical axis direction and two coordinates in the horizontal axis direction), and the section information is combined. As shown in 1102 of FIG. 11, the detection table after the combination of the section information is the token string $ (); [L% = $ (); [L% = N; $ (%,%,% ,%) Records the section information (9, 5)-(12, 7) for the file 9 and the section information (10, 5)-(13, 7) for the file 10. .

  Subsequently, the clone combining unit 104 executes the clone combining process in step S314 on the detection table illustrated in FIG. 11 in 1102. In the detection table illustrated in FIG. The beginning of the section information of the token sequence of the previous token string does not overlap, and the end of the section information of the preceding token string is not connected to the beginning of the section information of the subsequent token string. Therefore, in this clone combining process, the section information is not combined and the record is not registered in the new detection table. When the new detection table is not generated in this way, the process skips steps S316, S318, and S321, and proceeds to step S322.

  In step S322, the included clone deletion unit 105 extracts all the records in the detection table and adds them to the clone table. That is, all the records in the detection table indicated by 1102 in FIG. 11 are extracted and added to the clone table indicated by 1001 in FIG. As a result, the clone table shown in FIG. 12 is generated. In the clone table shown in FIG. 12, the final clone code detection result is registered. That is, token sequence $ (); [L% = N ;, $ (); [L% = $ (, $ (); [L% = $ (); [L% = N; $ (%,%, %,%) Clone code exists between files 9 and 10, and token string $ (); [L% = N; $ (); [L% = $ (clone code between files 11 and 12 The section information registered in the clone table is the longest matching section information by the section information combining process by the clone simple combining unit 103 and the clone combining unit 104.

  Subsequently, the included clone deletion unit 105 determines whether or not a new detection table has been generated (step S323). Since the new detection table has not been generated this time, the process ends.

  In the embodiment described above, the section information between the clone code detection target file and the file to be compared with the file (in this embodiment, the file at the time of the last clone code detection) is stored for each file. And the combined section information is registered in the clone table. Therefore, it is possible to achieve the longest match between the clone codes to be detected, and it is possible to prevent over-detection of the clone codes.

  In the example shown in FIG. 5, n-gram conversion is performed at n = 10. However, in actual processing, it is preferable to perform n-gram conversion at around n = 56 from the viewpoint of calculation efficiency. .

  Next, a second embodiment of the present invention will be described. In this embodiment, an example of evaluating the importance of a clone code among the clone codes detected in the first embodiment will be described. Specifically, the clone code detected in the first embodiment is subjected to clone analysis again in the same procedure, and the token sequence having the highest occupation rate (hereinafter referred to as a repeat region) is detected in the clone code. Identify region attributes. Then, the degree of importance of the repetitive region is specified from the specified attribute and occupancy rate, and it is determined whether or not to be excluded from the evaluation target. Details of the procedure and attributes will be described later.

  FIG. 13 is a block diagram showing a functional configuration of a clone code detection / evaluation apparatus 200 according to the second embodiment of the present invention. Compared to the block diagram shown in the first embodiment (FIG. 1), the block diagram shown in FIG. 13 is newly provided with a repeat area selecting unit 206, an attribute specifying unit 207, and a clone code analyzing unit 208. . In FIGS. 14 and 15 to be described later, a program that expresses it as definition information is stored in the ROM 203 or HD 207 in advance. The hardware configuration of the clone code detection / evaluation apparatus according to this embodiment is the same as the hardware configuration of the clone code detection / evaluation apparatus according to the first embodiment shown in FIG. Each of the repetitive region selection unit 206, the attribute specification unit 207, and the clone code analysis unit 208 has a configuration realized by a program stored in the HD 207 and loaded into the RAM 202 as necessary, and the CPU 201 that executes the program.

  FIG. 14 is a diagram showing the attributes of the repetitive region occupying the maximum region in the clone code detected in the first embodiment. The attribute of the repeat region is specified by “the length of the repeat region” and “the number of occurrences of the repeat region in the clone code detected in the first embodiment”.

  FIG. 15 is a diagram showing an evaluation map indicating the degree of importance. The degree of importance is specified by the “repetition region attribute” and the “repetition region ratio (cover rate) in the clone code detected in the first embodiment”. The ratio is “repetition region length” / “clone code length detected in the first embodiment”.

  The repetition area is detected by the repetition area selection unit 206 newly provided in the second embodiment. The detection procedure is the same as that in the first embodiment. In the first embodiment, the detection procedure is performed over a plurality of files. In the second embodiment, the “longest clone code detected in the first embodiment is used. Is performed.

  For example, when the “longest clone code (token string)” extracted in the first embodiment is “ABCDABCDCD”, the two-gram format “AB” “BC” “CD” “DA” “AB” “BC When “CD”, “DC”, and “CD” are combined, “ABCD” and “CD” are extracted. Next, the repetitive region selection unit 206 calculates the coverage of each of the extracted “ABCD” and “CD” in the “longest clone code”. That is, “ABCD / ABCDABCDCD = 80%” and “CD / ABCDABCDCD = 60%”. As described above, the token with the highest coverage is extracted as a “repetition area”. That is, “ABCD” occupying 80% of the ratio is the “repeated region”. If the coverage is the same, the shortest token is extracted as a “repetition area”.

  Next, the attribute specifying unit 207 obtains the “attribute” of the repeated area. Specifically, the attribute specifying unit 207 displays an attribute map corresponding to the ratio between the “repetition area length” and the “number of occurrences of the repetition area in the clone code detected in the first embodiment” shown in FIG. Based on the attribute map, the associated attribute is specified in accordance with the ratio of the “repetition area length” and the “appearance count” read from the ROM 203 or HD 207 and calculated by the repetition area selection unit 206. In the above example “ABCD”, “repetition region length” = 4 and “number of occurrences of the repetition region in the clone code detected in the first embodiment” = 2, so “(5) junk” It is.

  Next, the clone code analysis unit 208 reads an evaluation map corresponding to the ratio of “attribute” and “cover ratio” from the ROM 203 or the HD 207, and selects the “attribute” and repetition region of the repetition region identified by the attribute identification unit 207. According to the ratio of the “cover ratio” calculated in the unit 206, the associated importance level (“large”, “medium”, “small”, “N / A = not applicable”) is specified. In the above example “ABCD”, “attribute / junk” and “cover rate = 100% (1)” are “N / A”. FIG. 14 is an evaluation map assuming that the minimum token length is 100 or more. In actual evaluation, the specific ratio of “attribute” and “coverage” is appropriately changed according to the minimum token length and the like. Therefore, when the token length is 10 as in the above example, evaluation / analysis is performed with reference to an evaluation map having a specific ratio different from that in FIG.

  FIG. 16 is a diagram illustrating an example of a clone code determined to have an importance level of “small”. The clone code is only the field declaration part of Java (registered trademark). As a result of detection in the repeat area selection unit 206, the repeat area occupying the maximum area of the entire clone is one declaration of the field and the next private identifier ("private" "Object" "var" ";" "private") Since “repetition area length” = 5 and the number of repetition area appearances in the clone code detected in the first embodiment = 25, “repetition area attribute = Simple” “cover ratio = 25/25” = 1.00 ”and“ Importance = Small ”, and are deemed not to be evaluated.

  Actually, since the field declaration part is not a clone code of logic, no modification is necessary and the degree of importance that does not affect the quality of the source code is low. In other words, by excluding codes that do not need to be viewed as evaluation targets from the evaluation target, it is possible to reduce the number of analysts' work and improve efficiency.

  As described above, the embodiment of the clone code detection / evaluation apparatus has been described. However, the clone code detection / evaluation apparatus of the present invention is not limited to the clone code detection / evaluation apparatus having all the configuration requirements described in the above embodiment. Various changes and modifications are possible. Further, it goes without saying that such changes and modifications belong to the scope of the claims of the present invention.

  For example, FIG. 14 stored in advance in the ROM 203 or HD 207 described above is specified according to the ratio between the “repetition area length” and the “number of occurrences of the repetition area in the clone code detected in the first embodiment”. However, each specific numerical value does not need to be limited to FIG. The same applies to FIG. 15, which is specified according to the ratio of “attribute” and “cover ratio”, but each specific numerical value is not necessarily limited to FIG. 15.

100, 200 Clone detection device 101 Base table creation unit 102 Detection table creation unit 103 Clone simple combination unit 104 Clone combination unit 105 Included clone deletion unit 206 Repetitive region selection unit 207 Attribute specification unit 208 Clone code analysis unit

Claims (8)

Registers the segment information indicating the segment to the end position from the start position of each data of the first unit in clone code detection target file in the first detection table, and compared with the clone code detection target file First detection table generating means for registering, in the first detection table, section information indicating a section from a start position to an end position for each of the first unit data in the file,
Predetermined between the section information of the first unit data in the clone code detection target file registered in the first detection table and the section information of the first unit data in the comparison target file If conditions are satisfied, a binding means for coupling said first section information data of the first unit in the detection table for each of the files,
Section information combined for each of the files by the combining unit is extracted from the first detection table, and a second unit is formed by combining the data of the first unit corresponding to the section information combined by the combining unit. Clone code table generation means for extracting the clone code that is unit data and registering it in the clone code table;
A clone code detection / evaluation apparatus comprising: clone code evaluation means for evaluating the importance of the clone code based on a clone code registered in the clone code table . The clone code evaluation means specifies the attribute of the predetermined token sequence based on the length of the predetermined token sequence included in the clone code and the number of appearances of the predetermined token sequence in the clone code, 2. The clone code detection according to claim 1 , wherein the degree of importance of the clone code is evaluated based on the attribute of the specified predetermined token string and a ratio of the predetermined token string in the clone code. Evaluation device. The clone code detection / evaluation apparatus according to claim 2, wherein the predetermined token string is a token string having a highest ratio in the clone code among token strings extracted from the clone code. It said binding means includes a difference in section information between data of the first unit in the clones code detection target file, and the difference between the section information between data of the first unit in the comparison subject to file The clone according to any one of claims 1 to 3 , wherein, when the two are the same , the section information between the data of the first unit in the first detection table is combined for each file. Code detection evaluation device. Wherein any each file of the clones code detection target file and the comparison subject to file, the sintering said binding after treatment by the coupling means a first said in the detection table second unit of data section A second detection table generating means for registering, in the second detection table, section information obtained by combining section information of the second unit data for each arbitrary file when the information overlaps or is connected; Have
When the section information is registered in the second detection table by the second detection table generation unit, the combining unit is configured to output the second code in the clone code detection target file registered in the second detection table. When the predetermined condition is satisfied between the section information of the unit data and the section information of the second unit data in the file to be compared, the second unit in the second detection table The section information of data is combined for each arbitrary file, and the clone code table generating unit extracts the section information combined for each arbitrary file by the combining unit from the second detection table, and A third unit obtained by combining the data of the second unit corresponding to the section information combined by the second detection table generating unit It extracts clones encoding the data, clones encoding detection evaluation apparatus according to any one of claims 1 to 4, characterized in that registering the clones code table. Section of data of the third unit including section information of data of the second unit registered in the first detection table includes data of the second unit registered in the second detection table. If it is included in the information, further comprising a deletion means for deleting the section information of the data of the second unit from the first combined table,
The clone code table generation unit extracts, from the first detection table, section information excluding section information deleted by the deletion unit from the section information combined for each file by the combination unit. The clone code detection / evaluation apparatus according to claim 5 , wherein the clone code detection / evaluation apparatus is registered in a code table. A clone code detection evaluation method using a clone code detection evaluation apparatus,
It registers the segment information indicating the segment to the end position from the start position of each data of the first unit in clone code detection target file to detection table, in comparison with the file to which the said clone code detection target file A detection table generating step of registering section information indicating a section from a start position to an end position for each data of the first unit in the detection table;
Predetermined between the section information of the first unit data in the clone code detection target file registered in the first detection table and the section information of the first unit data in the comparison target file If the above condition is satisfied, a combining step of combining the section information of the data of the first unit in the detection table for each file;
The section information combined for each file by the combining step is extracted from the detection table, and the second unit data is formed by combining the first unit data corresponding to the section information combined by the detection step. A clone code table generation step for extracting the clone code and registering it in the clone code table;
A clone code evaluation step for evaluating the importance of the clone code based on a clone code registered in the clone code table . It registers the segment information indicating the segment to the end position from the start position of each data of the first unit in clone code detection target file to detection table, in comparison with the file to which the said clone code detection target file A detection table generating step of registering section information indicating a section from a start position to an end position for each data of the first unit in the detection table;
Predetermined between the section information of the first unit data in the clone code detection target file registered in the first detection table and the section information of the first unit data in the comparison target file If the above condition is satisfied, a combining step of combining the section information of the data of the first unit in the detection table for each file;
The second unit data obtained by extracting the section information combined for each file by the combining step from the detection table and combining the first unit data corresponding to the section information combined by the combining step. A clone code table generation step for extracting the clone code and registering it in the clone code table;
A program for causing a computer to execute a clone code evaluation step for evaluating the importance of the clone code based on a clone code registered in the clone code table .

Images