JP6235945B2 - Source code generation and provision apparatus, and source code generation and provision method - Google Patents

Description

  The present invention relates to a source code generation and provision apparatus and a source code generation and provision method.

  In system maintenance work, code (logic and data item definition code) is duplicated due to revision of source code over the years, deficiencies in design documents, or changes in data item and logic specifications, and the readability of source code There is a problem that the efficiency of maintenance work decreases. For this purpose, refactoring (reducing source code scale by duplicating common codes) has been performed manually, and maintenance has been restored.

  Patent Document 1 proposes a method for automatically generating source code using a script, and describes that an overlapping portion is made common when the same code as a generation result exists.

  Patent Document 2 describes that an existing code clone detection tool is used as a method for detecting duplicate codes, code clones are detected from a source file, and duplicate portions are shared.

JP 2007-304998 A JP 2009-176064 A

  In the method of sharing the duplicated code manually, there is a problem that the quality is low and the work cost is high. In addition, there is a problem in that it is difficult to detect similar source codes that are not completely matched, such as variable names being different, and that it is not possible to effectively duplicate code.

  Further, in the technique of Patent Document 1, if the means for sharing the duplicate code in the source code generation process, it can be said that the duplicate code that affects only the source code being generated is shared. There is a problem that it is not possible to share duplicate code that spans a plurality of programs because it is not possible to share duplicate code that spans source code that has already been generated at the time of conversion.

  The technology of Patent Document 2 is effective for a structured language in which the scope of variables can be localized. Variables such as COBOL85 are global variables (the entire program in which the variable influence range always declares variables). In the case of early structured programming languages, the problem is that generating variables included in duplicate code as function parameters makes the source code redundant and lowers the readability of the source code. . In addition, in the technique of Patent Document 2, the definition of the data structure composed of variable declarations is not an object of refactoring (code sharing).

  In addition, when the source code is corrected or maintained after the code is standardized, it is necessary to consider the change of the source code due to the code standardization, so that the maintainability after the code standardization is difficult to improve.

  An object of the present invention is to make it possible to efficiently generate refactored source code for source code created in an early structured programming language.

  Another object is to generate and provide maintenance information for facilitating maintenance work after refactoring.

The present invention analyzes a source code, detects similar partial codes included in the source code, and obtains position information on the source code, similar code detecting means,
A common code generating means for generating a common code based on the detected plurality of similar partial codes and for generating difference information between the common code and each of the similar partial codes;
Partial code replacement means for replacing each similar partial code in the source code with an insertion code for inserting difference information between the common code and the partial code based on the position information;
A source code generation device is provided.

The present invention provides the above-described source code generation device,
The similar partial code is a code clone which is a code considered to be equivalent in a program grammar,
The difference information is information for replacing a description in the partial code with another description, and provides a source code generation device.

The present invention provides the above-described source code generation device,
The similar code detecting means extracts a partial code similar as the code clone based on a program grammar, and compares a plurality of extracted partial codes or a part as a replacement target of the partial code by comparison with a reference. Provided is a source code generation device characterized by performing granularity adjustment by exclusion or division.

The present invention provides the above-described source code generation device,
A source code generation device comprising maintenance information generation means for generating maintenance information including the replaced similar partial code, the position information thereof, the common code, and the difference information I will provide a.

Further, according to the present invention, the similar code detecting unit analyzes the source code, detects mutually similar partial codes included in the source code, and acquires position information on the source code ; and
A common code generation unit generates a common code based on the detected plurality of similar partial codes, and generates difference information between the common code and each of the similar partial codes. When,
A partial code replacement step in which the partial code replacement means replaces each similar partial code in the source code with an insertion code for inserting the common code and the partial code difference information based on the position information. When,
A source code generation method characterized by comprising:

The present invention provides the above-described source code generation method,
The similar partial code is a code clone which is a code considered to be equivalent in a program grammar,
The difference information is information for replacing the description in the partial code with another description. A source code generation method is provided.

The present invention provides the above-described source code generation method,
In the similar code detection step, a partial code similar to the code clone is extracted based on the program grammar, and a part of the partial code to be replaced is compared by comparing the extracted partial codes or comparing with a reference. Provided is a source code generation method characterized by performing granularity adjustment by exclusion or division.

In addition to the source code generated in the above-described source code generation method, the present invention provides
A maintenance information providing step of providing maintenance information generated including the replaced similar partial code, the position information, the common code, and the difference information,
Furthermore, the present invention provides a source code providing method characterized by including the above.

  Unlike the conventional refactoring apparatus, the present invention extends the refactoring target not only to function units but also to variable declarations, so that more duplicate codes can be shared and maintainability can be further improved.

  Further, the present invention is a refactoring device suitable for source code created in an early structured language, and has a great effect when applied to a system with a long maintenance period that is expected to be used in an early structured language. Can demonstrate.

The figure which shows the form of this invention Example with a block. The figure which shows the procedure which performs the commonization of two similar partial codes, and the step which performs the automation. The figure which shows the Example of a procedure. The figure which shows the detail of the content of extraction of a similar code. The figure which shows the detail of the production | generation content of a common code. The figure which shows the detail of the content of replacement | exchange to a common code embedding instruction, and the content of the production | generation of the maintenance data. The figure which shows the example of the file 1, the file 2, and the file 3. The figure which shows the example of extraction of a similar code. The figure which shows the example of extraction of a similar code. The figure which shows similar partial code information. The figure which shows the production | generation of a common code. The figure which shows the replacement | exchange to a common code embedding instruction. The figure which shows that the embedding command was inserted about the file 1 and the file 2. FIG. The figure which shows the example of refactoring criteria information. The figure which shows the refactoring judgment example.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing a source code generating apparatus 100 according to an embodiment of the present invention.

  In FIG. 1, the source code generation device 100 refers to the source code 102 included in the processing device 101. It is assumed that the source code 102 is described in an initial structured program language in which variables are global variables.

  The source code generation device 100 includes a code clone detection unit 1, a code clone information acquisition unit 2, a sharing unit setting and code clone information adjustment unit 3, a code clone information acquisition unit 4, a partial code difference information generation unit 6, a sharing The code generation unit 7 includes an embedded code information acquisition unit 8, an adjustment source code generation unit 9, and a database 10. These means are connected to each other via a communication line 11 and can exchange information. The source code providing apparatus 110 includes a source code generating apparatus 100 and maintenance information generating means 105. The maintenance information generation means 105 is connected to the source code generation apparatus 100 and the communication line 13 so that information can be exchanged. The external processing device 101 and the source code providing device 110 are connected via a normal communication circuit 12. You may make it connect by another means. Mutual exchange of information generated by the communication circuit 12 is performed.

  When using a code clone, a similar partial code may be considered as a subordinate concept included in the code clone, but here the code clone (or clone) is related to a similar partial code with two similar partial codes Point to and use. When the partial codes are the same, they are shown as the same partial code.

  The code clone detection means 1 performs source code analysis and detection of similar partial codes and identical partial codes, which are partial codes grasped as code clones. The partial code is usually composed of the same partial code that is identical to each other and similar partial codes that are similar to each other, and is identified as a partial code. There is a limit to the common range in the common processing specified for the same partial code. In this embodiment, the common processing is also performed for similar partial codes. Although the similar partial code will be mainly described in this way, the common part common processing is not excluded. Various detection methods have been reported for detecting code clones, and various detection methods can be employed.

  The code clone detection means 1 analyzes a single or a plurality of source codes, and for at least two partial codes included in the source code, on the condition that the sequence of token types constituting the partial code is equivalent, Similar partial codes are detected as code clones. A token is a character string of a minimum unit constituting a source code in the grammar of a programming language.

  The code clone information acquisition unit 2 acquires position information on the source code of the partial code detected by the code clone detection unit 1. The position information may be referred to as area information.

  The code clone information adjustment unit 3 sets the common unit setting, reads the position information of the partial code acquired by the code clone information acquisition unit 2, and extracts a partial code having an arbitrary granularity (structural unit). . Granularity is set as a unit for code sharing. The granularity may be set to a plurality of granularities such as an arithmetic processing description and a data structure description. The granularity may be determined from the ease of understanding and maintainability of the source code generated by the source code generation device 100.

  The code clone information acquisition unit 4 acquires the position information on the source code of the extracted partial code with the common unit set by the code clone information adjustment unit 3. The partial code position information acquired by the code clone information acquisition means 2 is discarded.

  The partial code difference information generating means 6 detects a difference between partial codes. Differences are detected between partial codes having similar code relationships. Based on the detected difference, difference information between partial codes is acquired.

  The common code generation means 7 generates a common code according to a predetermined rule based on a predetermined unit for the code clone, typically a granularity. A rule that does not use the difference information between the partial codes for the generation of the common code may be defined. A difference between the generated common code and a partial code having a similar code relationship is detected. The detected difference information is acquired as difference information between partial codes. The partial code difference information acquired by the partial code difference information generation means 7 is discarded.

  The embedded code information acquisition unit 8 generates similar code information with difference information from the partial code difference information.

  The adjustment source code generation means 9 generates an instruction for deleting the partial code from each source code and a common code insertion instruction instruction for forming an embedded source code based on the processing result described above.

  The instruction to be deleted is an instruction to delete the partial code in relation to the position information of the acquired partial code, and the common code insertion instruction instruction to generate the embedded source code is inserted from the embedded code information to the deleted position. This is an instruction for embedding the generated common code and the generated difference information to generate an embedded source code without substantially changing the structural form of the source code. The embedded source code becomes the source code.

  In this way, embedded source code to be provided and used for the external processing apparatus 101 is generated. Thus, the generated slimmed source code 103 is used for processing in the processing apparatus 101.

  The source code providing apparatus 110 provides the processing apparatus 101 with the slimmed source code 103 and the maintenance information 104 generated by the generated source code maintenance information generating means 105. These two provisions make it easier for customers to use the source code.

  FIG. 2 shows a procedure including a procedure for sharing two similar partial codes and a procedure for automating the procedure.

In FIG. 2, the steps are roughly divided.
1. Similar code extraction step (i.e., code clone extraction step)
2. 2. Common code generation step 3. Replacement step to common code embedding instruction 4. Maintenance document generation step These steps comprise generation information provision steps, whereby the source code and related maintenance information are provided to the customer terminal of the user.
1. In the similar code extraction step, similar partial codes are extracted from the source code, and information on the similar partial codes is acquired.
1-1 Similar partial codes are extracted under the following conditions.
1) A unit for sharing. The unit of extraction is a predetermined arbitrary granularity as a unit to be shared 2) It is a syntax (code clone) regarded as equivalent in the program grammar 1-2 The following information is acquired as information of similar partial codes To do.
-Location information of similar partial codes on the source code In the common code generation step, a common code file is generated from the extracted similar partial codes.
2-1 Generate common code. When the common code is generated so that the difference with respect to the related similar partial code is reduced, the maintainability of the entire source code file is improved.
2-2 The difference between the generated common code and the related similar partial code is calculated.
2-3 Generate the generated common code as a common code file.
3. In the replacement step to the common code embedded instruction, replacement to the common code embedded instruction is performed. Each similar part code in the source code is deleted, and an instruction “insert the contents of the common code file while complementing the difference between the common code and the similar part code” is inserted at the deleted position (1-2a). . That is, a common code insertion instruction command is inserted. In this way, a slimmed source code having a deletion instruction and a common code insertion instruction instruction is generated.
4). In the maintenance material generation step, a list in which the relationships between the created common code and the corrected source code are summarized for each similar code is output as a file. Maintenance information is provided to customers along with the streamlined source code.

  FIG. 3 shows an embodiment of a source code providing procedure.

  In FIG. 3, the procedure is as follows: 1. extraction of similar partial codes and adjustment of codes; 2. Generation of common code 3. Replacement by inserting common code 4. Generation of maintenance materials. Providing generation information.

  In FIG. 3, source code 31 (source code 102 in FIG. 1) is prepared by extracting similar part codes, and the prepared source code 31 is referred to, and static analysis 32 (static analysis is conventionally known). Code clones containing similar partial codes are detected. The code clone information 33 cut out in this way and including the detected similar partial code is acquired. The code clone information 33 is referred to, and the code clone is adjusted, for example, the granularity adjustment 34 is performed. With reference to the refactoring granularity information 39, the code clone information 35 is acquired for the code clone for which the code clone granularity adjustment 34 has been made.

  The code clone becomes a refactoring candidate according to the refactoring adjustment criteria and is filed.

  With respect to the refactoring candidate code, the refactoring determination criterion information 36 is referred to, the refactoring determination 37 is performed, and the refactoring target code clone information 38 is generated.

  2. Move on to common code generation.

In this step, the difference 40 between the refactoring target code clones is detected 40. Difference information is generated from the detected difference. The code clone information 41 with difference information is acquired by referring to the difference information. The common code generation 42 is performed with reference to the code clone information 41 with difference information. From the generated common code and difference information, the common code 43 and common code information 44 for common processing are acquired.
3. Transition to replacement with common code embedding instructions.

For the transition, a delete instruction and a common code insertion instruction instruction are generated and inserted, that is, embedded 45 is made. Embedded source code 46 having this embedded instruction is generated and acquired. As a result, the size of the embedded source code is reduced as compared with the scale of the source code before maintenance without substantially changing the structural form of the source code before maintenance.
4). Generation of maintenance materials and 5. Shift to provision of generated information.

  In this step, maintenance information including source code processing method information is generated, and the generated source code information and maintenance information are provided.

In this way, it is a predetermined unit for analyzing a plurality of source codes of a single or a plurality of source programs to be executed by the processing device and sharing at least two partial codes included in the source code, And similar subcodes are detected, provided that the syntax is considered equivalent in program grammar,
Detects differences between code clones of multiple source codes and generates difference information for each source code,
For each source code, embedded code information consisting of the generated common code and the generated difference information is generated,
An instruction to delete the partial code in relation to the acquired position information and an embedded source code are generated, and the common code generated from the embedded code information and the generated difference information are embedded in the deleted position. Instructions are generated to form embedded source code without substantially changing the structural form of the source code,
Maintenance information on the relationship between each adjustment part code of the plurality of programs, the generated common code, and the embedded code information consisting of the detected difference is generated,
Provided is maintenance information on the relationship between the source code whose scale is reduced compared to the referenced source code and the embedded code information without substantially changing the structural form of the source code. As a result, the customer can perform various processes using the source code having the common code generated as described above.

  4 to 6 show details of the contents shown in FIG. The matter described in FIG. 3 will not be repeated more than necessary.

  FIG. The details of extraction of similar codes are shown. In FIG. 4, code clones are extracted by extracting similar partial codes. One code in the source code A is extracted from the source code A as the partial code (1) 32A, and becomes one partial code (1) 32B. Further, the similar code (1) 32C in the source code B is extracted from the source code B, and becomes a similar partial code (1) 32D. The two partial codes of one partial code (1) 32B and similar partial code (1) 32D are in a similar relationship and are in a code clone relationship.

  In this way, static analysis is performed, and code clone detection 32 for a plurality of source codes 31 is performed.

  Various methods for determining code similarity are known, and code similarity determination methods using known methods can be employed.

  Similar partial code information 33 is acquired from one partial code (1) 32B and similar partial code (1) 32D. Further, as the similar partial code information, the position of the similar code on the source code is acquired as the position information.

  In the next step, the refactoring granularity information 39 is referred to, and the code clone granularity adjustment 34 for adjusting the similar partial code to a common code unit is performed. The code clone information 35 in which the common code unit is adjusted to the code common execution unit is acquired. Adjust code clones to common code granularity. That is, it adjusts to a refactoring through particle size.

  The adjustment means that a part of the code of the similar partial code is deleted or the similar partial code is divided to match the code sharing execution unit, for example, the execution granularity. If the granularity is less than that of the common code, no sharing is required and it is excluded from the common processing. This adjustment improves the maintainability of the source code file after code sharing compared to the case where no adjustment is made.

  As for the adjustment unit and granularity unit, (1) program maintainability after common processing, such as functions and data structures, is not impaired, and (2) a larger granularity in the program language than the common code insertion instruction described later is set as appropriate. The That is, as the granularity, from the viewpoint of either or both of the arithmetic processing of the program and the data structure, the granularity on the program language structure is set as the refactoring granularity information in advance as the refactoring target.

  ◆ An example of cutting to a common granularity unit is shown. The code clone I34A is cut down to the granularity unit of the execution unit common processing, and is set as the code clone I34B. Therefore, the code clone I is a code clone that is set and adjusted in the unit of granularity of the common processing.

  ◆ An example of dividing into common granularity units is shown. The code clone II34C is divided up to the granularity unit of the execution unit common processing, and becomes a code clone II34D and a code clone II34E. Therefore, the code clone II34C and the code clone II34E become refactoring candidate codes that have been set and adjusted to a plurality of granularity units for common processing. The code clone is filed in a refactoring code clone information file as a refactoring candidate code.

  The refactoring candidate code is referred to the refactoring determination criterion information 36, so that the refactoring determination 37 (determination by the refactoring determination unit) is performed, the code clone information 38 is obtained as a refactoring target code clone. Hereinafter referred to as a code clone.

  The refactoring unit that performs the refactoring determination 37 includes the given determination information (configuration of source code to be refactored, etc.) and the calculated maintainability tendency (existing scale, for example, the scale and maintenance of the reduced source code). Based on the ease index, etc., refactoring candidate codes for refactoring were set from the target code clone information file.

  The determination is performed for each “description of data structure” and “description in operation” to be refactored. Then, as shown below, the difference information between the selected code clones is collected at the adjustment unit, for example, the token level granularity of the programming language, the result is linked to the previously selected code clone, and the refactoring target code clone is Acquire it and store it in the refactoring code clone information file.

  In this way, refactoring not only detects similar code as code clones, but also generates common code, collects difference information and uses the common code, and applies common code to create source code. Used in steps up to.

  FIG. Details of the contents of common code generation are shown.

  In the difference extraction 40 between code clones, a difference other than a common part between similar code clones is detected. There is no difference if they are exactly the same code, but there is a difference if they are similar. An embedding code including a common code and a difference related to the common code is generated.

  In a similar code clone, when a common part is common to both, and there is a code that is different only on one side, there is a case where a common part is common to both and there are codes that are different from each other. The common code generation is generated in the case where these differences exist, and when there is no difference, the difference is added to the common code when there is a difference.

  The code clone (1) 40A for the source code A and the code clone (1 ′) 40B for the source code B are code clones, each of which has a difference added. The code clone (1) 40A and the code clone (1 ′) 40B are identical portions between the code clones. Each difference is added to the same part in the similar part code. The code clone information 41 with difference information and the refactoring target code clone information 41 are acquired from the code clone (1) 40A and its difference, the code clone (1 ') 40A and its difference. This code clone information 41 with difference information is used as an embedding code.

  A common code 43 is generated from the code clone information 41, and a common code file is stored for the common code. The common code generated in this way is stored in the common code file.

  In the common code generation 42, a common code file is generated by extracting a common part from each code clone.

  The same code portion is made common among the code clones (1) and (1 ') of the source code A and other source codes. As a result, as shown by the common code 43, one or a plurality of common codes are generated corresponding to the target source code. When the common code is generated, the common code information 44 including the acquired code clone position information, the generated common code, and the difference between the detection target code clones is generated.

  FIG. 3. Details of replacement to common code embedding instruction and 4. Generation of maintenance materials. Indicates the contents of generation information provided.

  A common code insertion instruction instruction is generated, and instruction embedding 45 is performed.

  A partial code corresponding to the refactoring code clone is deleted from the program code, and an insertion instruction for the insertion code of each common code is inserted.

  Note that the common code to be inserted and the difference information from the common code are added to the common code insertion instruction.

  A partial code corresponding to the refactoring target code clone is deleted from the source code, and an insertion instruction for inserting each common code is generated, and insertion is instructed. Add difference information from the common code. The embedded source code 46 is generated.

  In FIG. 6, the partial code corresponding to the adjusted code clone is deleted from the position information 46A from the source code, and a common code insertion instruction command 46B for each generated common code is inserted. The common code insertion instruction embedding 46 is executed. Note that, in the common code insertion instruction, difference information from the common code is added to the common code to be inserted.

The embedded source code 46 becomes a source code generated as a result of common processing of the same partial code and similar partial code. The same partial code can be processed by the steps described above.
4). Maintenance material is generated, and 5, generation information is provided.

  In this maintenance material generation step, maintenance materials such as a common code list, a common code caller program list, and a deletion program list are generated together with the common processing.

The generated information can be provided.
・ Scale (number of code lines) list materials: List of scales such as the number of modified lines, number of deleted lines, number of added lines, etc. for each program source code file ・ Common code list: List of common codes ・ List of deleted codes: Common code List of modified code that shows the relationship of which similar code was replaced: List of parts that have been modified due to the effect of replacing similar code with common code When providing information, instructions for the generated source code Generate and store the embedded information as maintenance information for the original source code,
The external processing device is provided with source code whose scale is reduced compared to the referenced source code without substantially changing the structural form of the source code, and the maintenance information stored at the time of provision is provided. Provide as electronic information.

In this way, maintenance information of the relationship of the embedded code information consisting of each adjustment part code of the plurality of programs and the generated common code and the difference detected in relation to the common code is generated,
Source code providing for maintaining the relationship between the source code reduced in scale and the embedded code information compared with the referenced source code without substantially changing the structural form of the source code A device 110 is configured.

  7 to 14 show examples of common processing.

  FIG. 7 is an example of a COBOL record, and shows an example of file I, file 2, and file 3.

  8 and 9 show examples of similar code extraction.

  FIG. 8 shows an example of clone 1 and FIG. 9 shows an example of clone 2. No. of file 1 2-5, file 2 No. 2-5, file 3 No. 13-16 are coding clones.

  FIG. 15 and 16 and No19 and 20 are code clones.

  FIG. 10 shows similar partial code information.

  FIG. 11 shows the generation of the common code. No. 2-5 indicates the generated common code.

  FIG. 12 shows replacement to a common code insertion instruction command.

  The part with the background color (shaded) is corrected. No. of file 1 2-5, file 2 No. 2-5, file 3 No. 13-16 is logically deleted or physically deleted. The file 3 indicates that the common code insertion instruction is inserted into No17.

  It is executed by instructions that generate embedded source code without substantially changing the structural form of the source code.

  FIG. 13 shows that a common code insertion instruction command has been inserted for file 1 and file 2.

  FIG. 14 shows an example of refactoring determination criterion information.

  FIG. 15 shows an example of refactoring determination. Example 1 is a case where two data descriptions exist as code clone candidates but are not to be refactored. Example 2 is a case where two data descriptions exist as code clone candidates and are to be refactored.

The source code generation device 100 analyzes a plurality of source codes of a single or a plurality of source programs to be executed by a processing device, and uses at least a predetermined unit for sharing at least two partial codes included in the source code. A code clone detection means for detecting that the code clone is a similar partial code, on the condition that the syntax is considered to be equivalent in the program grammar;
For a plurality of source codes, code clone information acquisition means for acquiring that the two detected partial codes are in a code colony relationship and position information on the source codes of the partial codes,
Common code generation means for generating a common code for each code clone of a plurality of programs;
Adjustment source code difference information acquisition means for detecting differences between code clones of a plurality of source codes and generating difference information for each source code;
For each source code, embedded code information acquisition means for generating embedded code information composed of the generated common code and generated difference information;
An adjustment source code generation means including an instruction for deleting the partial code from each source code and an instruction for generating an embedded source code;
Here, the instruction to be deleted is an instruction to delete the partial code in relation to the acquired position information, and the instruction to generate the embedded source code is the common generated from the embedded code information at the deleted position. This is an instruction for embedding a code and generated difference information to generate an embedded source code without substantially changing the structural form of the source code.

The source code generation device 100
For each code clone of a plurality of source codes, a granularity adjustment setting unit that adjusts and sets the granularity as a unit for generating a shared code as a shared execution unit from the identified granularity unit according to a predetermined rule;
Code clone information acquisition means for acquiring information of the common code granularity adjusted and set as a common execution unit for each code clone of a plurality of source codes.

The common code generation means generates a common code for each code clone of a plurality of programs,
Difference information acquisition means between code clones detects differences between code clones of a plurality of source codes and generates difference information for each source code,
An embedded code information acquisition unit generates embedded code information including the common code and the generated difference information for each source code.

  The instruction to generate the embedded source code is embedded without substantially changing the structural form of the source code by embedding the common code and the difference generated in relation to the code clone from the embedded code information. This instruction generates source code.

  The code clone detecting means detects that the same and similar partial code relationship is found, and the code clone information acquiring means acquires information on the same and similar partial code. An instruction for generating an embedded source code is an instruction for generating a source code in which embedded code information having information indicating that there is no difference is also embedded.

  In this way, the same part code or / and similar part code of the program code can be shared, and the source code into which the above-described deletion and embedding instructions are inserted is formed, so that the scale of the source code is effective. Can be reduced.

  The structure of the source code does not change substantially, and maintenance is easy to use and maintainability. Therefore, it is versatile and can be shared efficiently. It is possible to provide a source code that has been scaled down and scaled down relative to the referenced source code.

In addition, source code created with a reduced scale can be provided.
And according to the present embodiment, between the common code granularity code clone-related partial code detected under the common code granularity set as described above, the similar partial code of the code clone relation and the generated common code Since the common code information consisting of the difference is acquired, for each source code, the similar partial code of the acquired code clone relation is deleted from the source code, and the generated common code and each detected common code granularity Code clone-embedded source code can be generated so that differences between code clone-like partial codes can be inserted. Further, since the same partial code or / and similar partial code of the source code can be shared under the set common code granularity, the scale of the source code can be effectively reduced.

  DESCRIPTION OF SYMBOLS 1 ... Code clone detection means, 2 ... Code clone information acquisition means, 3 ... Code clone information acquisition adjustment means, 4 ... Code clone information acquisition means, 6 ... Partial cone difference information generation means, 7 ... Common code generation means, 8 ... embedded code information acquisition means, 9 ... adjusted source code generating means, 10 ... database, 100 ... source code generating apparatus, 101 ... processing apparatus, 102 ... source code, 103 ... embedded source code, 104 ... maintenance information, 105... Maintenance information generating means, 110... Source code providing apparatus.

Claims (8)

Analyzing the source code, detecting similar partial codes included in the source code, and acquiring similar code detecting means for acquiring position information on the source code;
A common code generating means for generating a common code based on the detected plurality of similar partial codes and for generating difference information between the common code and each of the similar partial codes;
Partial code replacement means for replacing each similar partial code in the source code with an insertion code for inserting difference information between the common code and the partial code based on the position information;
A source code generation device comprising: In the source code generation device according to claim 1,
The similar partial code is a code clone which is a code considered to be equivalent in a program grammar,
The source information generating apparatus according to claim 1, wherein the difference information is information for replacing the description in the partial code with another description. In the source code generation device according to claim 2,
The similar code detecting means extracts a partial code similar as the code clone based on a program grammar, and compares a plurality of extracted partial codes or a part as a replacement target of the partial code by comparison with a reference. A source code generation device that performs granularity adjustment by exclusion or division. In the source code generation device according to any one of claims 1 to 3,
A source code generation device comprising maintenance information generation means for generating maintenance information including the replaced similar partial code, the position information thereof, the common code, and the difference information . A similar code detecting means analyzes the source code, detects similar partial codes included in the source code, and acquires position information on the source code ; and
A common code generation unit generates a common code based on the detected plurality of similar partial codes, and generates difference information between the common code and each of the similar partial codes. When,
A partial code replacement step in which the partial code replacement means replaces each similar partial code in the source code with an insertion code for inserting the common code and the partial code difference information based on the position information. When,
A source code generation method comprising: In the source code generation method according to claim 5,
The similar partial code is a code clone which is a code considered to be equivalent in a program grammar,
The source information generation method, wherein the difference information is information for replacing the description in the partial code with another description. The source code generation method according to claim 6,
In the similar code detection step, a partial code similar to the code clone is extracted based on the program grammar, and a part of the partial code to be replaced is compared by comparing the extracted partial codes or comparing with a reference. A source code generation method characterized by performing granularity adjustment by exclusion or division. In addition to the source code generated in the source code generation method according to claim 5,
A maintenance information providing step of providing maintenance information generated including the replaced similar partial code, the position information, the common code, and the difference information,
A source code providing method comprising:

Images