JP2024003945A - Program correction device, program correction method, and program correction program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and program for detecting and correcting a failure existing in a program with high accuracy.

SOLUTION: A program correction device 100 generates a single point correction candidate group 284 by eliminating a failure of a designated point of a source code of a program to be an object of software development, respectively generalizes (an abstraction failure point group 285, an abstraction source code 286) contents of the failure of the source code (the single point correction candidate group 284) in which the failure has been eliminated and the source code according to a prescribed rule, estimates another failure point (an additional failure point group 287) in the source code on the basis of the generalized contents of the failure, and applies processing for creating information (an additional material code group 288) for eliminating the other failure point and the created information to the other failure point in the source code on the basis of the failure of the designated point and the source code.

SELECTED DRAWING: Figure 2

COPYRIGHT: (C)2024,JPO&INPIT

Description

The present invention relates to a program modification device, a program modification method, and a program modification program.

In software development, when a malfunction occurs in a program (an operation that differs from the developer's intention), the program is modified so that the malfunction does not occur. The causes of program failures are called defects (defects, faults, or bugs). Debugging is the process of identifying a location in a program's source file that contains a defect and rewriting the location to remove the defect. For debugging, test files that represent the behavior intended by the developer are often used.

Developers manually debug, but this has the problem of taking a long time. Therefore, Non-Patent Document 1 discloses a technique for automating debugging work. Specifically, the technique described in Non-Patent Document 1 is a technique for correcting the defect by identifying a location containing a defect from a source file (fault localization) and rewriting the location (patch generation). This technique follows a fix pattern to change a code fragment containing a defect to a correct source code fragment (donor) that exists elsewhere in the source file to be debugged.

Seemanta Saha, Ripon K Saha, and Mukul R Prasad. Harnessing evolution for multi-hunk program repair. In Proceedings of the 41st International Conference on Software Engineering, pages 13-24. IEEE Press, 2019.

The technique described in Non-Patent Document 1 detects a plurality of similar locations and rewrites them at the same time, thereby correcting a problem that requires rewriting of multiple locations. However, since the detection of similar parts is performed line by line in the source file, if there are multiple defective parts and they are approximated only in part of the line, it is difficult to detect and correct the multiple defective parts at the same time. I can't. In addition, if a location that is calculated to have a low probability of containing a defect is included among multiple defective locations, other similar locations will not be detected, so all defective locations will be detected. cannot be rewritten correctly. Furthermore, since the modification history of the source file is used to detect similar parts, it cannot be applied to software that does not have a modification history.

The present invention has been made in view of the above circumstances, and its purpose is to provide a program correction device, a program correction method, and a program correction method that can accurately detect and correct defects existing in a program. The purpose is to provide programs.

One of the present inventions for solving the above problems is a storage device for storing a program, a location of a defect in the program to identify the defect, and a content of the defect in the program in which the defect has been identified. and a process of generalizing the program according to a predetermined rule, and estimating the location of another defect in the program based on the content of the generalized defect, and determining the location of the other defect based on the identified defect and the program. The present invention is a program correction device that includes a processing device that executes a process of creating information for resolving the location of the problem, and a process of applying the created information to the other defective location in the program.

According to the present invention, defects existing in a program can be detected and corrected with high accuracy.
Structures, effects, etc. other than those described above will be made clear by the description of the embodiments below.

FIG. 2 is a diagram illustrating an example of a hardware configuration of a program modification device. FIG. 2 is a diagram illustrating an example of functions included in the program modification device. FIG. 3 is a diagram showing an example of a source file group described in this embodiment. FIG. 3 is a diagram showing an example of a test file group described in this embodiment. 3 is a flowchart illustrating an example of program correction processing. FIG. 3 is a diagram showing an example of a defective location candidate. FIG. 3 is a diagram illustrating an example of a group of single location correction candidates. FIG. 3 is a diagram illustrating an example of an abstracted defect location group. FIG. 3 is a diagram illustrating an example of abstracted source code. FIG. 7 is a diagram illustrating an example of a group of additional defective locations. It is a figure which shows an example of an additional material code group. FIG. 7 is a diagram illustrating an example of a group of multiple correction candidates. It is a figure which shows an example of a correction result.

This embodiment will be described below with reference to the drawings. In the following description, the same or similar configurations may be denoted by the same reference numerals and redundant description may be omitted.

FIG. 1 is a diagram showing an example of the hardware configuration of the program modification device 100.
The program modification device 100 is an information processing device including a processor 110, a main storage device 120, an auxiliary storage device 130, an input device 140, an output device 150, and a communication device 160. These devices are communicably connected to each other via a communication means such as a bus (not shown).

The processor 110 is, for example, a CPU (Central Processing Unit) or an MPU (Micro Processing Unit).
This is a processing device configured using a processing unit. When the processor 110 reads and executes programs stored in the main storage device 120, various functional units of the program modification device 100, which will be described later, are realized.

The main storage device 120 is a device that stores programs and data, and is, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), or a nonvolatile semiconductor memory (NVRAM (Non Volatile RAM)). The main storage device 120 stores each program described below.

The auxiliary storage device 130 is, for example, a hard disk drive, SSD (Solid State Drive), optical storage device (CD (Compact Disc), DVD (Digital Versatile Disc), etc.).
, a storage system, an IC card, a reading/writing device for a recording medium such as an SD memory card or an optical recording medium, or a storage area of a cloud server. Auxiliary storage device 13
Programs and data stored in 0 are read into the main storage device 120 at any time.

The input device 140 is, for example, a keyboard, a mouse, a touch panel, a card reader, a voice input device, or the like, and receives input of various information from the user. The output device 150 is a user interface that provides the user with various information such as processing progress and processing results, and includes, for example, a screen display device (liquid crystal monitor, LCD (Liquid Crystal Display), graphic card, etc.), an audio output device (speaker, etc.). etc.), or a printing device, etc.

Note that instead of the input device 140 or the output device 150, the program modification device 100 may be configured to input or output information to or from another input/output device connected via the communication device 160.

The communication device 160 is a wired or wireless communication interface that realizes communication with other devices via a communication network such as a LAN (Local Area Network) or the Internet, and is, for example, a NIC (Network Interface Card). , a wireless communication module, a USB (Universal Serial Interface) module, or a serial communication module.

Here, terms used in this embodiment will be defined.
"Source code" is a program that is the target of software development, and the description language and data format are not particularly limited. It is assumed that the source code of this embodiment (before modification by the program modification device 100) includes a plurality of defective parts that cause an operation different from the developer's intention.

A "method" is a unit of processing in a program. A method includes one or more instructions (Statement). An instruction includes zero or more functions or expressions.
Instructions and expressions together are called a "code fragment." The source code includes one or more such methods.

A "field" is a variable or constant that can be referenced from all methods. The source code includes zero or more fields.

A "method call" refers to an expression (function) that performs a call process to call a method.

"Field access" refers to an expression (function) that performs reference processing that refers to a field.

"Scope" is information that specifies (limits) the access range (call range or reference range) of a method or field when calling a method or referencing a field in a source code. For example, a method call is expressed in the format "scope.method(parameter)", and a field access is expressed in the format "scope.field".

"Test code" is a program that tests the operation of source code. The test code is a program written to verify that the program performs the operations intended by the developer.

A "test method" is a unit of processing in a test code. For example, a test method is a program that tests the operation of one method. Note that a test method that tests a certain method is called a "corresponding test method" for that method. One method may have two or more corresponding test methods. The test code includes one or more such test methods.

Next, FIG. 2 is a diagram showing an example of the functions included in the program modification device 100. The program modification device 100 of this embodiment attempts to modify a program by rewriting a code fragment in a source code containing a plurality of defects into a material code fragment. At this time, if you can rewrite the correct material code piece, you can modify the program.

As shown in the figure, the program correction device 100 includes a defect location identification section 210, a single location correction candidate generation section 220, an additional defect location extraction section 230, an additional material code generation section 240, a multiple location correction candidate generation section 250, It includes functional units such as a correction candidate inspection unit 260 and an information storage unit 280.

The defective location identification unit 210 receives the source file group 281 and the test file group 282 as input, and outputs a defective location candidate group 283 (see FIG. 6 for details).

The single location correction candidate generation unit 220 inputs the source file group 281 and the defective location candidate group 283 and outputs a single location correction candidate group 284 (see FIG. 7 for details).

The additional defect location extracting unit 230 receives the single location correction candidate group 284 as input and outputs an abstracted defect location group 285 (see FIG. 8 for details).

Further, the additional defective part extracting unit 230 inputs the source file group 281 and the single part correction candidate group 284 and outputs 286 (see FIG. 9 for details).

Then, the additional defective part extracting unit 230 receives the abstracted defective part group 285 and the abstracted source code 286 as input, and outputs an additional defective part group 287 (see FIG. 10 for details).

The additional material code generation unit 240 inputs the source file group 281 and the single location correction candidate group 284 and outputs the additional material code group 288 (see FIG. 11 for details).

The multiple location correction candidate generation unit 250 inputs the single location correction candidate group 284, the additional defective location group 287, and the additional material code group 288, and outputs a multiple location correction candidate group 289 (see FIG. 12 for details). .

The modification candidate inspection unit 260 inputs the test file group 282, the single location modification candidate group 284, and the multiple location modification candidate group 289, and outputs a modification result 291 (see FIG. 13 for details).

The information storage unit 280 includes a source file group 281, a test file group 282, a defect location candidate group 283, a single location correction candidate group 284, an abstracted defect location group 285, an abstracted source code 286, and additional files. A group of defective locations 287, a group of additional material codes 288, a group of multiple location correction candidates 289, and correction results 291 are stored.

The source file group 281 is a collection of source files whose problems are to be resolved. A source file consists of one or more lines. The language of the source code written in the source file is not particularly limited, and may be, for example, Java (registered trademark) or C++.

The test file group 282 includes one or more test code files (test files) for verifying whether the source files of the source file group 281 have a defect. A test file consists of one or more lines.

(source files)
Here, FIG. 3 is a diagram showing an example of the source file group 281 described in this embodiment. This source file group 281 includes a source file 310 with the file name File1 and a source file 320 with the file name File2.

The source file 310 with the file name File1 contains nine lines of source code, and lines 1 to 9
The contents of the method named checkCubeAndCalcVolume are described in the first line.

It is assumed that the source code of this source file 310 includes defects in the third and fifth lines. Specifically, in order to resolve this problem, change "SHAPES.REC" (code 311) in "String[ ] keys={SHAPES.REC, SHAPES.TRI, SHAPES.HEX};" in the third line. Rewrite it to "SHAPES.CUBE" and "SHAPES.REC" in "if(shape.is(SHAPES.REC)){" on the 5th line.
” (code 312) must be rewritten to “SHAPES.CUBE”.

(Test file group)
Next, FIG. 4 is a diagram showing an example of the test file group 282 described in this embodiment. The test file group 282 includes a test file 410 whose file name is File1Test.

This test file 410 has 11 lines of test code. The test file 410 has a test method named testVolumeCube1 in lines 1 to 5, and a test method named testVolumeCube1 in lines 7 to 1.
The first line has a test method named testVolumeCube2. Test methods testVolumeCube1 and testVolumeCube2 are both corresponding test methods of method checkCubeAndcalcVolume in source file 310.

Note that the program developer can change the side length by using the method checkCubeAndCalcVolume in the source file 310 for lines 2 to 4 and lines 8 to 10 of the test code in the test file 410 when 0 is input. The intention is to output the volume 8 of the cube of 2, and when 2 is input, output "-1" indicating that it is not a cube.

Next, the defect location candidate group 283 shown in FIG. 2 is information that specifies the position (line in this embodiment) of a defect location in the source code.

The single location correction candidate group 284 is a set of source files (hereinafter referred to as first correction candidates) in which one defective location (code fragment) in the source code is rewritten.

The abstract defect location group 285 is a code fragment (hereinafter referred to as an abstraction) that is an abstraction (generalization) of a code fragment having a defect (hereinafter referred to as a corrected code fragment) that was rewritten in the single location correction candidate group 284. This is a collection of modified code fragments).

The abstracted source code 286 is a collection of source codes obtained by abstracting (generalizing) the source code by rewriting part of the source file (code fragment) with a predetermined symbol (placeholder). In this embodiment, the placeholder is assumed to be "$ID", but other character strings may be used.

The additional defective portion group 287 is a collection of information for identifying lines in the source code that include defective portions other than corrected code fragments (hereinafter referred to as additional corrected code fragments).

The additional material code group 288 corrects each defective part in the single part correction candidate group 284 (
This is a collection of code fragments used for (rewriting). Note that, hereinafter, each code piece in the additional material code group 288 will be referred to as a material code piece.

The multiple location correction candidate group 289 is a collection of information on source codes (hereinafter referred to as second correction candidates) in which defective locations (code pieces) have been rewritten using each material code piece of the additional material code group 288.

The correction result 291 is information on the source code in which the problem has been resolved.

The functions of each functional unit of the program modification device 100 described above are realized by the processor 110 reading a program created corresponding to each functional unit from the main storage device 120 or the auxiliary storage device 130. These programs can be distributed by being recorded on, for example, a portable or fixed recording medium. It should be noted that each function of the program modification device 100 may be implemented in whole or in part by a virtual server provided using virtualization technology, process space separation technology, etc., such as a virtual server provided by a cloud system. It may also be realized using information processing resources. Further, all or part of each functional unit of the program modification device 100 may be realized by, for example, a service provided by a cloud system via an API (Application Programming Interface) or the like.
Next, the processing performed by the program modification device 100 will be explained.

<Program correction process>
FIG. 5 is a flowchart illustrating an example of a program modification process performed by the program modification device 100. The program modification process is started, for example, when the program modification device 100 receives an instruction input from a user, or by batch processing that is set to be executed at a scheduled timing. In addition, in the following, the letter "S" added before the reference numeral means a processing step.

First, the defect location identification unit 210 reads the source file group 281 and the test file group 282 stored in the information storage unit 280, executes the test file, and obtains information on the identified defect location in the source file group 281. It is stored in the defect location candidate group 283 (S111).

The problem is, for example, a method call to a scoped method or a field access to a scoped field. The defect location may be identified using, for example, static analysis or fault localization. Note that the defect location specifying unit 210 may definitively resolve the defect identified in S111.

(Problem location candidates)
Here, FIG. 6 is a diagram showing an example of the defect location candidate group 283.

The defect location candidate group 283 includes one or more entries. The entry of the problem location candidate group 283 includes columns of file name 2831, line number 2832, and suspicion 2833. The file name 2831 stores the file name of a source file that may contain a defect, and the line number 2832 stores the line number of the defective location in the source file. Suspicion 2833 stores the probability that the defective location contains a defect. In the example shown in the figure, the entry 2834 of the defect location candidate group 283 records that there is a high possibility (1.0) that the description in the third line of the source code has a defect location.

Note that if the location containing the defect can be uniquely identified, the entry does not necessarily need to have information set in all of the file name 2831, line number 2832, and suspicion level 2833.

Next, as shown in FIG. 5, the single location correction candidate generation unit 220 reads the source file group 281 and defective location candidate group 283 stored in the information storage unit 280 and analyzes them. A single location correction candidate group 284 including one correction candidate is generated (S112).

Specifically, the single location correction candidate generation unit 220 selects one location from among the defect locations registered in the defect location candidate group 283, and rewrites the selected location with another correct description. A group of one-place correction candidates 284 is generated. The single location correction candidate group 284 may be generated using a correction template, for example, or may be generated by brute force (Brute-force).
However, it may also be created manually by a developer or the like.

(Single location correction candidates)
FIG. 7 is a diagram showing an example of the single location correction candidate group 284. This single location modification candidate group 284 includes one or more single location modification candidates 2841 (first modification candidates). The single location correction candidate 2841 stores the source code 310 of the source file group 281 in which one of the defect locations registered in the defect location candidate group 283 has been rewritten.

In the example shown in the figure, the source code of the single location correction candidate 2841 is to change the field access 311 "SHAPES.REC", which is the defective location on the third line of the source file 310 (File 1), to the field access 2842 "SHAPES.REC". This is the source code rewritten to “SHAPES.CUBE”.

Next, as shown in FIG. 5, the additional defect location extraction unit 230 performs the following processes P1 to P3 based on the source file group 281 and the single location correction candidate group 284 stored in the information storage unit 280. Accordingly, the additional failure location group 287 is output (S113).

Process P1: The additional defect location extracting unit 230 abstracts the code fragments in the single location correction candidate group 284 and outputs it to the abstracted defect location group 285.

Specifically, the additional defect extraction unit 230 extracts the specified part of a method in a method call whose scope (calling range) is specified and the scope (reference range) specified in the single-location correction candidate group 284. Each reference part of a field in field access is abstracted by replacing it with a placeholder "$ID", and this is output to the abstraction defect location group 285. On the other hand, if the code fragments in the single location correction candidate group 284 are neither method calls nor field accesses, the additional defect location extraction unit 230 outputs them as is to the abstracted defect location group 285 without replacing them with placeholders. . Note that if another method call or field access appears as a parameter in a method call, the additional defect extraction unit 230 may replace only the innermost method call or field access with the placeholder "$ID". good.

Further, the additional defect extraction unit 230 does not need to replace with placeholders the method specification part in a method call whose scope is not specified and the field reference part in a field access whose scope is not specified. Thereby, by reducing the number of abstracted code fragments, it is possible to suppress output of unnecessary additional defect points in process P3, which will be described later.

(Abstract defect location group)
FIG. 8 is a diagram showing an example of the abstracted defect location group 285.

The abstracted defective location group 285 includes one or more abstracted defective locations 2851. The abstract defective location 2851 is a code fragment that is an abstraction of a code fragment (corrected code fragment) whose description has been rewritten in the single location correction candidate group 284, that is, information that includes an abstracted corrected code fragment.

For example, the field access 311 "SHAPES.REC" of the source file 310, which is a code fragment of the single location correction candidate group 284 rewritten in the process of S112, is a code fragment with a clear scope, and is specifically has scope “SHAPES” and field “REC
"including. Therefore, the additional defect extraction unit 230 abstracts the modified code fragment 2852 by rewriting the field "REC" part of the field access 311 "SHAPES.REC" to the placeholder "$ID". “SHAPES.$ID” is output to the abstract defect location group 285.

Process P2: Next, the additional defect extraction unit 230 abstracts the entire source code included in the source file group 281 (regardless of whether there is a defect there) and outputs it to the abstracted source code 286. . The method of abstraction is the same as the process P1. However, the additional defect location extracting unit 230 does not abstract lines that have been rewritten using the single location correction candidate group 284.

(abstraction source code)
FIG. 9 is a diagram showing an example of the abstract source code 286.

This abstracted source code 286 is source code information that is abstracted by rewriting a part of the source code included in the source file 310 in the source file group 281 to a placeholder "$ID".

In the figure, the additional defect extraction unit 230 selects the source file 310, which has been output (rewritten) to the single location correction candidate group 284 through the process of S112, out of the source code included in the source file group 281. By abstracting all abstractable codes other than the third line 2861 of "File 1", an abstracted source code 286 is output.

For example, the method call 313 "shape.is(SHAPES.REC)" on the fifth line of the source file 310 "File1" has a field access 312 "SHAPES.REC" for the parameter (this is rewritten with the single location correction candidate group 284). ), the additional defect extraction unit 230 outputs this as it is to the abstracted source code 286 without abstracting it. Additionally, the method call 314 "Math.power(shape.length1,3)" on the 6th line of the source file 310 "File1" includes a field access 315 "shape.length1" whose scope is clearly specified in the parameter, so there is an additional problem. The location extraction unit 230 abstracts this and outputs it to the abstracted source code 286 (reference numeral 2862).

Process P3: Next, the additional defect extraction unit 230 selects a line that includes an abstracted correction code piece registered in the abstracted defect group 285 from among each line of the abstracted source code 286, and , in the process of S112, all lines that are different from the line for which the code fragment was rewritten (that is, lines containing additional defects) are identified by the single location correction candidate group 284, and the information of each identified line is added to the additional defect location group 287. Output.

(Additional defective locations)
FIG. 10 is a diagram showing an example of the additional defect location group 287.

The additional failure location group 287 includes zero or more entries. The entry of the additional defect location group 287 includes columns of file name 2871 and line number 2872. The file name 2871 stores the file names of the source files 310 and 320 that include the additional defect, and the line number 2872 stores the line number of the line that includes the additional defect.

In the example shown in the figure, the additional defect extraction unit 230 extracts code fragments registered in the abstracted defect group 285 from among each line of the abstracted source code 286 that are abstracted from the source file 310 (File 1). The fifth line 286 contains the code fragment 2852 “SHAPES.$ID”.
3 is associated with the source file 310 (File 1) and output to line number 2872 of the additional defect location group 287.

Next, the additional material code generation unit 240 generates an additional material code group 288 based on the single location correction candidate group 284 (S114).

Specifically, for example, if the code fragment (modified code fragment) rewritten in the single location modification candidate group 284 is a field access, the additional material code generation unit 240 generates the same code fragment from the source file group 281. Other field access expressions that refer to fields included in the scope are extracted, and the field portion is output to the additional material code group 288. Further, for example, when the modified code fragment is a method call, the additional material code generation unit 240 extracts from the source file group 281 an expression for another method call that calls a method included in the same scope as the code fragment. Then, the additional material code generation unit 240 excludes method invocation expressions having a different number or type of parameters from the code fragment from the extracted method invocation expressions, and outputs the method invocation part to the additional material code group 288. do.

Note that if no code piece is used to generate the single location correction candidate group 284, the additional material code generation unit 240 outputs an additional material code of 0.

(Additional material code group)
FIG. 11 is a diagram showing an example of the additional material code group 288. The additional material code group 288 includes zero or more additional material codes 2881.

The additional material code 2881 is, for example, a code fragment used for rewriting in the single location correction candidate group 284, a field access expression that refers to another field included in the same scope as the code fragment, or a code fragment that is used for rewriting the single location correction candidate group 284. This is a method call expression that calls another method included in the same scope. The additional material code 2881 shown in the figure includes a field access 2881a "SAHPES.CUBE" and field access 2881b, c ("SHAPES.CUBE") which are other code fragments that refer to other fields included in the same scope as that code fragment. .TRI”, “SHAPES.HEX”).

Note that the additional material code group 288 does not need to be set with the code pieces (corrected code pieces) rewritten in the single-location correction candidate group 284 . For example, in the additional material code group 288 shown in the figure, the code piece "SHAPES.REC" that was rewritten in the single location correction candidate group 284 is not set.

Next, as shown in FIG. 5, the multiple location correction candidate generation unit 250 generates information on the second correction candidate based on the single location correction candidate group 284, the additional defective location group 287, and the additional material code group 288. A group of multiple correction candidates 289 is output (S115).

Specifically, the multiple location correction candidate generation unit 250 refers to the abstract source code 286 for each code piece of the location (line) indicated by the additional defect location group 287 from the source code of the single location correction candidate group 284. By rewriting the defective parts of each identified code piece with each material code piece registered in the additional material code group 288, a plurality of correction candidate group 289 is output. Note that the multiple location correction candidate generation unit 250 can perform rewriting using the same method as for generating the single location correction candidate group 284, for example, but the rewriting method is not particularly limited.

(Group of multiple correction candidates)
FIG. 12 is a diagram showing an example of a group of multiple correction candidates 289. The multiple location correction candidate group 289 includes zero or more multiple location correction candidates 510 (510a, 510b, 510c, . . . ) (second correction candidates). The multiple location correction candidates 510 correct defective locations in each line registered in the additional defective location group 287 in the source code of the single location correction candidate group 284 using each material code piece of the additional material code group 288. Contains rewritten source code.

The multiple location correction candidate 510a shown in the figure includes the field access “SHAPES.REC” at the location indicated by line number 2872 of the additional defective location group 287 (line 5 of File 1), and the addition of the additional material code group 288. This is the source code rewritten using the field access 2881a "SHAPES.CUBE" registered in the material code 2881. In addition, multiple correction candidates 51
0b is the location indicated by line number 2872 of the additional defect location group 287 (5 of File 1).
This is a source code in which the field access “SHAPES.REC” in line 288 is rewritten with another field access 2881b “SHAPES.TRI” registered in the additional material code 2881 of the additional material code group 288. In addition, the multiple location correction candidate 510c uses the field access "SHAPES.REC" of the location indicated by line number 2872 of the additional defect location group 287 (fifth line of File 1) to the additional material code of the additional material code group 288. This is the source code rewritten with yet another field access 2881c "SHAPES.HEX" registered in 2881.

Next, as shown in FIG. 5, the modification candidate inspection unit 260 uses the test file group 282, the single location modification candidate group 284, and the multiple location modification candidate group 289 stored in the information storage unit 280. , outputs the correction result 291 (S116).

Specifically, the modification candidate inspection unit 260 tests each of the modification candidates (first modification candidate, second modification candidate) included in each source code of the single location modification candidate group 284 and the multiple location modification candidate group 289. By inputting data to each test file of the file group 282, output data of each test file is obtained. If the output data of the test file indicates that the line including the defective part has been correctly rewritten, the modification candidate inspection unit 260 outputs the modification candidate as the modification result 291. Note that static analysis may be used, for example, or a test code may be used to check whether the line including the defective part has been correctly rewritten.

Thereafter, the program modification device 100 may repeat the process of S112 to create a new single location modification candidate group 284 (S117). At this time, the program correction device 100 includes an additional defect location extraction unit 230, an additional material code generation unit 240, a multiple location correction candidate generation unit 250, an abstracted defect location group 285, an abstracted source code 286, an additional defect location group 287, The additional material code group 288 and the like may also be initialized. As a result, independent first correction candidates can be generated for a plurality of first correction candidates, and more defects can be removed without affecting each other.

At this time, the single location modification candidate generation unit 220 of the program modification device 100 may further select other entries in the defective location candidate group 283. As a result, if a problem cannot be solved by changing the code piece in a certain line, it is possible to try changing the code piece in another line and solve the problem with a higher probability.

(Result of correction)
Here, FIG. 13 is a diagram showing an example of the correction result 291. The modification result 291 shown in the same figure includes field access 311 "SHAPES.REC" in the third line of the source code of the source file 310 to field access 2911 "SHAPES.CUBE" (first modification candidate), and Field access 312 "SHAPES.REC" Field access 2912 "SHAPES.CUBE"
” (second correction candidate), the source code in which the problem was resolved by rewriting each is stored.

As described above, the program correction apparatus 100 of the present embodiment eliminates defects at specified locations in the source code, generates the single location correction candidate group 284, and generates the source code (single location correction The content of defects in the candidate group 284) and the source code are generalized according to predetermined rules (abstracted defect location group 285, abstracted source code 286), and based on the generalized content of defects, other defects in the source code are generalized. The location was estimated (additional defect location group 287), and based on the single location correction candidate group 284 and the source code, information for resolving other defect locations was created (additional material code group 288). Apply the information to other defective locations in the source code.

That is, the program correction device 100 can broadly estimate other defective parts of the source code and create correction candidates by abstracting the initially identified defective part and the source code and referring to both. Other defects in the source code can be corrected using the created correction candidates.

In this way, according to the program correction apparatus 100 of the present embodiment, it is possible to accurately detect and correct defects existing in a program. This makes it possible to make the developer's debugging work more efficient.

Furthermore, the program modification apparatus 100 of this embodiment generalizes the source code by replacing each of the method call method and field access field whose scope is specified with a placeholder.

This makes it possible to identify locations in the source code where bugs commonly occur, and to efficiently modify the source code.

Furthermore, the program modification apparatus 100 of this embodiment treats field and method portions in field accesses and method calls whose scopes are not explicitly specified as part of the generalized source code as they are.

In this way, by not performing abstraction processing when abstraction is not necessary, it is possible to perform processing for identifying and resolving defective locations at higher speed.

In addition, the program correction device 100 of this embodiment eliminates defects in the scoped method call portion or the field access portion of the scoped field (single location correction candidate group 284), and abstracts the defect location group. 285 to generate part of another method call within the same scope, or a field access within the same scope.

As a result, method calls or field accesses that call or produce methods or fields within the same scope can be used as material for fixing other defects, and points that are likely to have defects can be efficiently fixed. Can be fixed.

The present invention is not limited to the embodiments described above, and can be implemented using arbitrary components without departing from the spirit thereof. The embodiments and modifications described above are merely examples, and the present invention is not limited to these contents as long as the characteristics of the invention are not impaired. Furthermore, although various embodiments and modifications have been described above, the present invention is not limited to these. Other embodiments considered within the technical spirit of the present invention are also included within the scope of the present invention.

For example, some of the hardware included in each device of this embodiment may be provided in another device.

Further, each program of the program modification device 100 may be provided in another device, a certain program may be made up of a plurality of programs, or a plurality of programs may be integrated into one program.

100 Program correction device 210 Defect location identification unit 220 Single location correction candidate generation unit 230 Additional defect location extraction unit 240 Additional material code generation unit 250 Multiple location correction candidate generation unit 260 Correction candidate inspection unit

Claims (6)

a storage device that stores the program; and
The location of the defect in the program is identified, the contents of the defect in the program in which the defect has been identified and the program are generalized according to predetermined rules, and based on the contents of the generalized defect, other defects in the program are determined. A process of estimating the location,
A process of creating information for eliminating the other defective parts based on the identified defect and the program;
A program correction device comprising: a processing device that executes a process of applying the created information to the other defective location in the program. The processing device includes:
A specified part of a process that calls another process for which a call range is specified, or a part that refers to a constant or variable in a process that refers to a constant or variable for which a reference range is specified. generalizing the program by replacing with the predetermined symbol;
A program modification device according to claim 1. The processing device is a specified part of a process that calls another process for which a call range is not specified, or a constant in a process part that refers to a constant or variable for which a reference range is not specified. 3. The program modification device according to claim 2, wherein the reference portion of the variable is made a part of the generalized program as it is. The processing device includes:
Identifying defects in the part of the call process that calls other processes for which the call range is specified, or the part of the reference process that refers to constants or variables for which the reference range is specified,
As information for resolving the other problems, a part of the calling process that calls other processes within the calling range, or a part of the reference process that refers to other constants or variables within the reference range, Create information to resolve the other problems mentioned above;
A program modification device according to claim 1. The information processing device
memorize the program,
The location of the defect in the program is identified, the contents of the defect in the program in which the defect has been identified and the program are generalized according to predetermined rules, and based on the contents of the generalized defect, other defects in the program are determined. A process of estimating the location,
A process of creating information for eliminating the other defective parts based on the identified defect and the program;
applying the created information to the other defective location in the program;
How to modify the program. In the information processing device,
memorize the program,
The location of the defect in the program is identified, the contents of the defect in the program in which the defect has been identified and the program are generalized according to predetermined rules, and based on the contents of the generalized defect, other defects in the program are determined. A process of estimating the location,
A process of creating information for eliminating the other defective parts based on the identified defect and the program;
executing a process of applying the created information to the other defective location in the program;
Program fix.

Images