JP2016164727A - Test case selection device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a test case selection device with which it is possible to reduce the number of test cases selected as necessary to reexecute in a regression test and improve the efficiency of a process for testing a program.SOLUTION: The test case selection device comprises: storage units 102, 103 for storing a program before and after alteration; a storage unit 104 for storing a test case and a statement executed in a program before alteration as execution information; a storage unit 105 for storing an expected value that is a variable outputted in a test case; an acquisition unit 106 for obtaining different statements in the program and after alternation as a difference statement; an analysis unit 107 for obtaining a statement affected from the difference statement as a dependency analysis result; and a selection unit 109 for selecting a test case that has the possibility of the value of an expected value being changed in the program and after alteration, as a test case executed in a regression test, using the statement of execution information, the expected value, the difference statement, and the dependency analysis result.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a test case selection device that can select a test case that needs to be re-executed in a regression test of a program.

  When the program is modified, a test for confirming whether an unintended influence other than the modified portion has spread, that is, whether the program is not defective due to the modification is called a regression test. In the simplest embodiment of regression testing, all prepared test cases are retested. However, in this case, the larger the software scale, the more man-hours are required to retest all the test cases, and the efficiency of the regression test decreases. Also, if you postpone retesting for individual corrections that are repeated until product release because of the large number of test steps, it is likely that important defects will be overlooked until late in development. For this reason, regression test selection is generally performed that reduces the number of regression test steps by selecting and executing only existing test cases that require retesting.

  As an example of the regression test selection method, there is a method of recording in advance which statement of a program is executed by a test case and automatically selecting a test case for executing a corrected portion. Patent Document 1 describes an example of a regression test selection method using a similar method. In the system and method described in Patent Document 1, a statement related to the corrected location is extracted according to priority (strength of dependency), and a test case associated with the extracted statement is executed. , Streamlining regression testing.

JP 2008-204405 A

  In view of the fact that regression testing tends to be very large in large-scale software development, it is desirable that the regression test selection method reduce the number of test cases to be retested as much as possible. In particular, it is essential to reduce the number of test cases when performing regression tests for each program modification separately from tests for checking the final quality of software.

  However, in the method of selecting a test case that executes a correction part or a part related to the correction (a part affected by the correction) as in the technique described in Patent Document 1, a test case that cannot find a defect in the program is used. There is a problem of selecting a large number. For example, when there is a correction at a location where many test cases are executed in common, such as in the initialization processing of a program, the number of test cases executed in the regression test can hardly be reduced. In addition, most of the test cases can be executed, except for those testing cases that execute such modified parts, except those that test the initialization process itself or parts that are affected by the modified initialization process. This is a test case in which a bug in the program cannot be found.

  As described above, in the conventional regression test selection method, the number of test cases executed in the regression test may not be reduced. In such a case, there is a problem that the efficiency of the test process of the program is lowered.

  An object of the present invention is to provide a test case selection device capable of reducing the number of test cases to be selected as being necessary for re-execution during a regression test and improving the efficiency of a program test process. To do.

  The test case selection apparatus according to the present invention has the following features. The variable that is output when the test case is executed and the value of this variable is the expected value, the expected value is determined for each test case, and the program before the change and the change that has been changed to the program before the change A storage unit for storing a post-program, a test case execution information storage unit for storing the test case executed for the pre-change program and a statement executed in the executed test case as test case execution information; , A test case expected value information storage unit for storing the expected value variable as test case expected value information for each test case, and a program difference for obtaining a different statement as a difference statement in the pre-change program and the post-change program The acquisition section and the statement affected by the difference statement Comprising a program dependency analysis unit for determining the presence relationship analysis, and a test case selector which selects the test case to be executed during regression testing of the post-change program. The test case selection unit uses the dependency analysis result from the statement of the test case execution information, the expected value of the test case expected value information, and the difference statement, and the program before change A test case that may change the value of the expected value variable with the changed program is obtained, and the obtained test case is selected as the test case to be executed in the regression test.

  According to the present invention, it is possible to provide a test case selection device that can reduce the number of test cases that are selected as needing re-execution during a regression test, and can improve the efficiency of the test process of the program. .

The figure which shows an example of a structure of the test case selection apparatus by the Example of this invention. The figure which shows an example of the program before a change stored in the program storage part before a change. The figure which shows an example of the program after a change stored in the program storage part after a change. The figure which shows an example of the test case execution information stored in the test case execution information storage part. The figure which shows an example of the test case expected value information stored in the test case expected value information storage part. The figure which shows an example of the process sequence which a program difference acquisition part performs. The figure which shows an example of the process sequence which a program dependence analysis part performs. The figure which shows an example of the dependence relationship analysis result stored in the dependence analysis result storage part. The dependence analysis result graph which visualized the dependence shown in FIG. The figure which shows an example of the process sequence which a test case selection part performs. The figure which shows an example of the hardware constitutions of a test case selection apparatus.

  The test case selection device according to the present invention needs to be re-executed by preventing a test case in which a defect caused by the correction of the program cannot be found from being selected as a test case that needs to be re-executed during the regression test of the program. Reduce the number of test cases to choose as and increase the efficiency of regression testing.

  Here, we will describe in detail what is a test case worth re-execution, that is, a test case that can detect defects caused by program modifications.

  The test case is executed to check whether the program performs the expected operation, the procedure for executing the program (statements executed in the test case and the order in which they are executed), and when executing the procedure Is defined by the input variables and their values necessary for the output, and the output variables and their values (expected values) obtained when the procedure is executed. “Execute a test case” means to execute a statement defined in the test case. Usually, a plurality of test cases are defined in advance.

  The variables and their values that are output when the test case is executed are called “expected values”. Expected values are determined for each test case. That is, an expected value is set according to the test case.

  Therefore, a malfunction caused by the modification of the program is found when there is a difference between the execution result of the test case, that is, the specific program output and the expected value determined for the test case. From this, a test case in which a defect can be found is a test case in which the effect of correction affects the program output, and as a result, a difference occurs between the program output and the expected value corresponding thereto.

  One reason for selecting a large number of test cases in which a defect cannot be found by the conventional method is that the information about the expected value defined for the test case is not used.

  In the test case selection device according to the present invention, information on the expected value determined in the test case is used, and whether or not the influence of the modification of the program reaches the program output corresponding to the expected value (the program output is changed by the modification of the program). Whether or not there is a possibility that it is different from the expected value) is used as a judgment criterion when selecting a test case. Thus, in the test case selection apparatus according to the present invention, during the regression test of a program, a test case in which a defect caused by the modification of the program cannot be found is not selected as a test case that needs to be re-executed. Increase efficiency.

  Hereinafter, a test case selection apparatus according to an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram illustrating an example of a configuration of a test case selection apparatus 101 according to an embodiment of the present invention. The test case selection apparatus 101 includes a pre-change program storage unit 102, a post-change program storage unit 103, a test case execution information storage unit 104, a test case expected value information storage unit 105, a program difference acquisition unit 106, and a program dependence analysis unit 107. , A dependency analysis result storage unit 108, a test case selection unit 109, and a user interface (I / F) unit 110. The test case selection apparatus 101 can be configured by a computer, for example.

  The pre-change program storage unit 102 stores program source code (pre-change program) before being changed (corrected). The post-change program storage unit 103 stores the program source code (the post-change program) that has been changed as part of the pre-change program. The test case selection apparatus 101 inputs the pre-change program and the post-change program, and stores the pre-change program in the pre-change program storage unit 102 and the post-change program in the post-change program storage unit 103, respectively.

  The test case execution information storage unit 104 stores test cases executed for the pre-change program and statements executed in the test cases as test case execution information T. The test case execution information T is obtained by executing each test case on the pre-change program in advance and is stored in advance in the test case execution information storage unit 104.

The test case expected value information storage unit 105 stores, as test case expected value information E, the expected value determined for each test case and the program output information related to this expected value. The test case expected value information E is information obtained when each test case is defined, and is stored in the test case expected value information storage unit 105 in advance.
The program difference acquisition unit 106 inputs a pre-change program and a post-change program, and compares and analyzes these programs to obtain a difference statement. The difference statement will be described later with reference to FIG.

  The program dependency relationship analysis unit 107 inputs a difference statement, obtains a statement having a dependency relationship with the difference statement, and analyzes the dependency relationship. A statement having a dependency relationship and an analysis result thereof are stored as a dependency relationship analysis result G in the dependency analysis result storage unit 108. The dependency relationship will be described later with reference to FIG.

  The test case selection unit 109 selects a test case that may change the value of the expected value variable between the pre-change program and the post-change program as a test case to be executed during the regression test of the post-change program. Specifically, the test case selection unit 109 receives the test case execution information T, the test case expected value information E, and the dependency relationship analysis result G, and the influence of the program modification reaches the program output corresponding to the expected value. A test case that needs to be re-executed is selected based on the determination of whether or not the program output may be different from the expected value due to the modification of the program.

  The user interface unit 110 provides a means for transmitting processing results of the test case selection apparatus 101 and instructions from the user 111. The user interface unit 110 can be configured by, for example, an input device such as a computer mouse or keyboard, an output device such as a display or a printer, and a program for controlling them.

  FIG. 2 is a diagram illustrating an example of the pre-change program stored in the pre-change program storage unit 102. FIG. 2 shows a part of a program source code written in C language as an example. In this embodiment, each line of the program source code is referred to as a “statement”. Each statement is assigned a statement number (1) to (13).

  FIG. 3 is a diagram illustrating an example of the changed program stored in the changed program storage unit 103. The post-change program is stored in the post-change program storage unit 103 in the same format as the pre-change program shown in FIG. 2, but a part of the pre-change program is changed. The statement with the statement number (1) is changed, the statement with the statement number (5a) is added, and the statement with the statement number (12) in the pre-change program is deleted. Hereinafter, “statement with statement number (1)” is abbreviated as “statement (1)”, and statements with other statement numbers are also abbreviated in the same manner.

  In FIG. 2 and FIG. 3, the symbol “*” is added to the changed part of the program source code. That is, the symbols “*” are attached to the statements (1), (5a), and (12).

  FIG. 4 is a diagram illustrating an example of the test case execution information T stored in the test case execution information storage unit 104. The test case execution information T is information on the statement number of the statement executed when each test case is executed for the pre-change program. As shown in FIG. 4, each test case is given a “test case No.” so as to be uniquely identifiable, and a statement number (“executed statement information”) executed in the test case is assigned. It is associated. That is, the test case execution information T is information in which a test case executed for the pre-change program is associated with a statement executed in the test case.

  FIG. 5 is a diagram illustrating an example of the test case expected value information E stored in the test case expected value information storage unit 105. The test case expected value information E is information on the program output (expected value variable) corresponding to the expected value determined for each test case. As shown in FIG. 5, each test case is associated with a program output (expected value variable) corresponding to the expected value of the test case. This program output can be obtained in advance based on the expected value set when each test case is defined. In addition to variables, the program output can include the RAM area address of the program, parameters to be given to the function, the return value of the function, and the like.

  In the test case expected value information E shown in FIG. 5, there are two variables z and h related to the program shown in FIGS. Considering only these two variables, test case no. 1-3 show the output of the variable z, the test case no. 4 to 6 can be said to be test cases for examining the output of the variable h.

  FIG. 6 is a diagram illustrating an example of a processing procedure executed by the program difference acquisition unit 106.

  In process 601, the pre-change program P is input from the pre-change program storage unit 102, and the post-change program P ′ is input from the post-change program storage unit 103.

  In the process 602, the pre-change program P and the post-change program P ′ are compared, and different statements between them are obtained as difference statements to obtain a difference statement set S that is a set of difference statements. The difference statement is a statement that is different between the pre-change program P and the post-change program P ′, and thus is a statement that has been changed or deleted in the pre-change program P and a statement that has been added to the post-change program P ′. It is. In the present embodiment, the modified statement (1) shown in FIGS. 2 and 3, the deleted statement (12) shown in FIG. 2, and the added statement (5a) shown in FIG. It is a difference statement. That is, the set of statements with the symbol “*” in FIGS. 2 and 3 is the difference statement set S.

  It should be noted that a statement corresponding to the before-change program P and the after-change program P ′ obtained when the processing 602 is executed is checked later. In this embodiment, this correspondence between statements can be seen from the statement number. In FIG. 2 and FIG. 3, statements having the same statement number are corresponding statements. Further, the number of the statement added to the post-change program P ′ (FIG. 3) is a statement (addition) corresponding to the added statement among the statements of the pre-change program P (FIG. 2), such as “5a”. The letter “a” is added after the number “5” of the statement one line before the generated statement so that the correspondence can be understood.

  The process 603 returns the difference statement set S obtained in the process 602 as an input of the subsequent process.

  FIG. 7 is a diagram illustrating an example of a processing procedure executed by the program dependency analysis unit 107.

  In process 701, the pre-change program P, the post-change program P ′, and the difference statement set S are input from the program difference acquisition unit 106.

  In the process 702, the statements of the pre-change program P and the post-change program P ′ are analyzed, and a statement that is dependent on the statements (difference statements) of the difference statement set S is obtained. The statements (dependency source statement and dependency destination statement) in the obtained dependency relationship are output as the dependency relationship analysis result G together with the dependency type. As will be described later, the types of statement dependency are at least “data dependency” and “control dependency”. That is, the program dependence analysis unit 107 obtains the dependence analysis result G by examining at least “data dependence” and “control dependence”.

  A dependency is a relationship in which, in two statements, one is changed, deleted, or added, and the other is affected. Dependencies include not only direct dependencies from difference statements, but also indirect dependencies through other statements. That is, as a statement having a dependency relationship, for example, not only a statement including a variable described in a difference statement (a statement having a direct dependency relationship) but also from a difference statement without including a variable described in a difference statement Also ask for affected statements (statements with indirect dependencies). The dependency relationship analysis result G includes only the dependency relationship that can be traced from the difference statement in the flow of statement execution in the program.

  In process 703, the dependency relationship analysis result G obtained in process 702 is stored in the dependency analysis result storage unit 108. Further, the difference statement set S input from the program difference acquisition unit 106 is stored in the dependency analysis result storage unit 108.

  FIG. 8 is a diagram illustrating an example of the dependency relationship analysis result G stored in the dependency analysis result storage unit 108. As shown in FIG. 8, in the dependency relationship analysis result G, the dependency source statement, the dependency destination statement, and the type of dependency are associated with each other as a dependency relationship.

  In the dependency type, “data dependency” means that there is an influence (dependency relationship) due to a variable. For example, in the pre-change program shown in FIG. 2, since the change to the variable x in the statement (1) affects the statement (3) using the variable x, there is no data between these statements. There is a dependency relationship.

  In the type of dependency, “control dependency” means that there is an influence (dependency relationship) due to a conditional branch such as an if statement. For example, in the pre-change program shown in FIG. 2, since the statement (8) controls whether or not the statement (9) is executed, there is a control-dependent relationship between these statements.

  The dependency analysis result storage unit 108 stores the dependency analysis result G by omitting the redundant dependency relationship between the pre-change program P and the post-change program P ′. For example, as shown in FIGS. 2 and 3, the dependency relationship from the statement (1) to the statement (3) is duplicated between the pre-change program P and the post-change program P ′, and thus one of them is omitted.

  FIG. 9 is a dependency analysis result graph in which the dependency relationship shown in FIG. 8 is visualized and shown in an easy-to-understand manner. In FIG. 9, square boxes with rounded corners indicate statements, and arrows indicate dependency relationships. The statement at the tail side (tail) of the arrow is the dependency source statement, and the statement at the head side (head) of the arrow is the dependency destination statement. Statements have a statement number. The characters shown beside the arrow indicate the type of dependence, “D” indicates data dependence, and “C” indicates control dependence. A statement with the symbol “*” added to the statement number is a difference statement.

  FIG. 10 is a diagram illustrating an example of a processing procedure executed by the test case selection unit 109.

  In process 1001, the difference statement set S and the dependency analysis result G are input from the dependency analysis result storage unit 108, the test case execution information T is input from the test case execution information storage unit 104, and the test case expected value information E is tested. Input from the case expected value information storage unit 105.

  In process 1002, one test case t is selected from the test case execution information T (FIG. 4). It is assumed that all test cases included in the test case execution information T can be selected without duplication. In the following description, test case No. A case where 1 (see FIG. 4) is selected will be described as an example.

  In process 1003, a statement set S 'is obtained. The statement set S ′ is a set of statements included in the difference statement set S executed in the test case t. That is, the statement set S ′ is a difference statement included in the test case t.

  For example, the test case t is a test case No. In the case of 1, the test case no. The statements executed by 1 are statements (1), (2), (3), (6), (7), (8), (9) and (13) as shown in FIG. Since the set of statements with the symbol “*” in FIGS. 2 and 3 is the difference statement set S, the difference statement set S includes statements (1), (5a), and (12). Therefore, the statement set S ′ consists of the statement (1). The statement (5a) added to the post-change program P ′ (FIG. 3) is a case where the statement (5) corresponding to the statement (5a) is executed among the statements of the pre-change program P (FIG. 2). It shall be executed.

  In process 1004, it is determined whether or not the statement set S 'is empty. If the statement set S ′ is empty, the test case t does not execute the difference statement, and the process proceeds to process 1010. If the statement set S ′ is not empty, the process proceeds to processing 1005.

  In process 1005, a program output (variable) in the test case t is extracted from the test case expected value information E (FIG. 5), and an expected value information set e is created. Test case t is test case no. In the case of 1, the variable z is extracted from the test case expected value information E. Therefore, the expected value information set e consists of the variable z.

In process 1006, obtain a set S _'2 comprising a statement affected by the statements contained in the' statement and the statement set S contained in the 'statement set S. The statement affected by the statement included in the statement set S ′ is a statement that depends directly or indirectly on the statement included in the statement set S ′, and the dependency analysis result G (FIG. 8) and the dependency analysis result. It is obtained using the graph (FIG. 9). Test case t is test case no. In the case of 1, the statement set S ′ consists of statements (1), so the set S ′ ₂ contains statements (1), (3), (8), (9), (11), and (12). Consists of. This can be easily obtained by following the arrow indicating the dependency relationship from the statement (1) in the dependency analysis result graph shown in FIG.

In process 1007, 'in the _2, a set of statements S for changing the value of a variable included in the expected value information set e' set S get '. A statement that changes the value of a variable is a statement that directly changes the value of a variable included in the expected value information set e by an operator such as an assignment operator “=” and an increment operator “++” in C language. It is. Test case t is test case no. In the case of 1, the set S ′ ₂ includes statements (1), (3), (8), (9), (11), and (12), and the variable included in the expected value information set e is z. Of the statements (1), (3), (8), (9), (11), and (12), the statement (9) changes the value of the variable z. (9).

  In process 1008, when the test case t is executed, any one of the statements included in the statement set S ″ is affected by the statements included in the statement set S ′ (a set of difference statements included in the test case t). Find out whether or not to receive. That is, whether or not any one statement in the statement set S ″ directly or indirectly depends on (ie, is affected by) the difference statement executed by the test case t when the test case t is executed. Find out. At this time, it is possible to examine whether or not it is affected (whether or not it depends) using the dependency relationship analysis result G (FIG. 8) and the dependency analysis result graph (FIG. 9).

  Test case t is test case no. In the case of 1, the statement set S ″ includes the statement (9), and the statement set S ′ includes the statement (1). For example, when the dependency analysis result graph (FIG. 9) is used, the statement (9) can be traced from the statement (1) via the statements (3) and (8) to the statement (9). ) To be affected via statements (3) and (8) (statement (9) depends indirectly on statement (1) via statements (3) and (8)). Test case t is test case no. In the case of 1, statements (1), (2), (3), (6), (7), (8), (9) and (13) are executed (FIG. 4). Therefore, test case No. Executing 1 makes it possible to trace from statement (1) to statement (9), ie, statement (9) is affected by statement (1) (depending on statement (1)).

  When it is determined in process 1008 that any one of the statements included in the statement set S ″ is affected by the statement included in the statement set S ′ when the test case t is executed, the process 1009 is performed. If it is determined that there is no influence, the process proceeds to the process 1010.

  In the processing 1008, when the dependency relationship analysis result G (FIG. 8) and the dependency analysis result graph (FIG. 9) are used to determine whether or not they are affected (whether they depend), the dependency of the statement If the type of is "control dependent", it is considered that the dependency destination statement is executed unconditionally after executing the dependency source statement. This is for considering the possibility that the transition destination of the conditional branch changes due to the influence of the modification of the program. For example, if the statement (8) is executed, it is considered that the statements (9), (11), and (12), which are the control-dependent dependencies of the statement (8), are also executed.

  In process 1009, the test case t is selected as a test case that needs to be re-executed. In the above example, the test case No. 1 is selected as a test case that needs to be re-executed.

  In process 1010, it is determined whether or not all test cases included in the test case execution information T have been examined. If all the test cases have been examined, the process proceeds to process 1011. If not, the process proceeds to process 1002.

  In process 1011, the test case selected in process 1009 is output to the user interface unit 110 as a selection result. The user interface unit 110 can display the selection result on an output device such as a display or a printer.

  In the above example, in process 1002, the test case No. The case where 1 is selected has been described. Next, in process 1002, the test case No. No. 2 is selected and test case no. The case where 5 is selected will be briefly described.

  In the process 1002, the test case No. When 2 is selected, it is as follows.

Test case No. As shown in FIG. 4, the statements executed by 2 are statements (1), (2), (4), (5), (6), (7), (8), (9) and (13). It is. In the process 1003, a statement set S ′ composed of difference statements (1) and (5a) is obtained. Since the statement set S ′ is not empty, the process proceeds to processing 1005 to obtain an expected value information set e including the variable z. In the process 1006, statements (1), (3), (5a), (8) are set as a set S ′ ₂ composed of statements included in the statement set S ′ and statements affected by the statements included in the statement set S ′. ), (9), (11) and (12). In process 1007, 'among _2, a set of statements S changing the value of the variable z included in the expected value information set e' set S as' to obtain a set consisting of the statement (9).

  In the process 1008, the test case No. When 2 is executed, it is checked whether the statement (9) included in the statement set S ″ is affected by the statement (1) or (5a) included in the statement set S ′. From the dependency analysis result graph shown in FIG. 9, the statement (9) is not affected by the statement (5a), and the statement (9) is transferred to the statement (1) via the statements (3) and (8). It can be seen that it is affected. As shown in FIG. Since 2 does not execute statement (3), statement (9) is not affected by statement (1). Therefore, test case No. When executing 2, statement (9) is not affected by either statement (1) or (5a). Therefore, the process 1009 is not executed (the test case No. 2 is not selected as a test case that needs to be re-executed), and the process transitions to the process 1010.

  In the process 1002, the test case No. When 5 is selected, it is as follows.

Test case No. As shown in FIG. 4, the statements executed by 5 are statements (1), (2), (4), (5), (6), (7), (8), (11), (12) And (13). In the process 1003, a statement set S ′ composed of difference statements (1), (5a) and (12) is obtained. Since the statement set S ′ is not empty, the process transits to a process 1005 to obtain an expected value information set e including the variable h. In the process 1006, statements (1), (3), (5a), (8) are set as a set S ′ ₂ composed of statements included in the statement set S ′ and statements affected by the statements included in the statement set S ′. ), (9), (11) and (12). In process 1007, 'in the _2, a set of statements S for changing the value of the variable h included in the expected value information set e' set S as' to obtain a set consisting of the statement (11).

  In the process 1008, the test case No. 5 is executed, whether or not the statement (11) included in the statement set S ″ is affected by the statement (1), (5a), or (12) included in the statement set S ′. From the dependency analysis result graph shown in FIG. 9, the statement (11) is indirectly influenced from the statement (1) through the statements (3) and (8) and directly from the statement (5a). It turns out that it is influenced. As shown in FIG. 5 does not execute statement (3). When executing 5, statement (11) is not affected by statement (1). However, test case no. 5 executes statement (5), so test case no. 5 is executed, the statement (5a) is executed, and the statement (11) is affected by the statement (5a) (as described above, the statement (5a) is changed to the statement ( If statement (5) corresponding to 5a) is executed, it shall be executed). Therefore, test case No. When executing 5, statement (11) is affected by statement (5a). Therefore, in process 1009, the test case No. 5 is selected as a test case that needs to be re-executed, and the process proceeds to the process 1010.

  The test case selection apparatus according to the present embodiment performs the same processing as described above with No. 1 shown in FIG. 1-No. No. 6 test case, test case no. 1, no. 4 and no. 5 is selected as a test case that needs to be re-executed. Therefore, the test case selection apparatus according to the present embodiment can reduce the number of test cases that are selected as needing re-execution in the regression test, and can improve the efficiency of the test process of the program. On the other hand, in the case of a method for selecting a test case for executing a modified part of a program, which has been conventionally used, No. 1-No. Since all of the six test cases execute the corrected statement (1), the number of test cases to be re-executed cannot be reduced.

  FIG. 11 is a diagram illustrating an example of a hardware configuration of the test case selection apparatus 101. As illustrated in FIG. The test case selection apparatus 101 can be configured by a computer, for example, and includes a CPU 1102, a main storage device 1103, an input device 1104, an output device 1105, an auxiliary storage device 1106, and an optical drive 1107, which are connected via a bus 1101. Connected to each other.

  As the input device 1104, for example, both or one of a mouse and a keyboard can be used. As the output device 1105, for example, a display and / or a printer can be used.

  The optical drive 1107 reads a test case selection program 1108 recorded on the optical disc 1111. The test case selection program 1108 is a program for operating the test case selection device 101 and is stored in the auxiliary storage device 1106.

  The auxiliary storage device 1106 includes a test case selection program 1108, an operation system (OS) 1110 that is an operating environment of the test case selection program 1108, a pre-change program P, a post-change program P ′, test case execution information T, and a test. Development data 1109 such as case expected value information E is stored. The development data 1109 can be input by the input device 1104.

  The CPU 1102 loads the test case selection program 1108 stored in the auxiliary storage device 1106 to the main storage device 1103 and executes it. When the CPU 1102 executes the test case selection program 1108, the test case selection apparatus 101 operates.

  A series of operations for executing the test case selection program 1108 is input from the user by the input device 1104. A series of processing results executed by the test case selection program 1108 is output by the output device 1105 and transmitted to the user.

  As described above, the test case selection apparatus according to the present embodiment determines whether or not the influence of the program modification reaches the program output corresponding to the expected value determined in the test case (the program output is changed to the expected value by the program modification). Whether or not there is a possibility of difference) is used as a criterion for selecting a test case. For this reason, the test case selection apparatus according to the present embodiment can reduce the number of test cases to be re-executed during the regression test and improve the efficiency of the test process of the program.

  The test case selection device according to the present invention can be used for selecting a test case that needs to be re-executed in a regression test for an arbitrary program. For example, a control system for automobiles, railways, elevators, etc. It can be used for other programs.

  In addition, this invention is not limited to said Example, A various deformation | transformation is possible. For example, the above-described embodiments are described in detail for easy understanding of the present invention, and the present invention is not necessarily limited to an aspect including all the configurations described. Further, it is possible to delete a part of the configuration of the embodiment or to add or replace another configuration.

  DESCRIPTION OF SYMBOLS 101 ... Test case selection apparatus, 102 ... Program storage part before change, 103 ... Program storage part after change, 104 ... Test case execution information storage part, 105 ... Test case expected value information storage part, 106 ... Program difference acquisition part, 107 ... Program dependency analysis unit, 108 ... Dependence analysis result storage unit, 109 ... Test case selection unit, 110 ... User interface unit, 111 ... User, 1101 ... Bus, 1102 ... CPU, 1103 ... Main storage device, 1104 ... Input device DESCRIPTION OF SYMBOLS 1105 ... Output device, 1106 ... Auxiliary storage device, 1107 ... Optical drive, 1108 ... Test case selection program, 1109 ... Development data, 1110 ... Operation system, 1111 ... Optical disk.

Claims (4)

The expected value is the variable output when the test case is executed and the value of this variable. The expected value is determined for each test case.
A storage unit for storing a pre-change program and a post-change program changed to the pre-change program;
A test case execution information storage unit for storing the test case executed for the pre-change program and a statement executed in the executed test case as test case execution information;
For each test case, a test case expected value information storage unit that stores the expected value variable as test case expected value information;
A program difference acquisition unit for obtaining a different statement as a difference statement between the pre-change program and the post-change program;
A program dependency analysis unit that obtains a statement affected by the difference statement as a dependency analysis result;
A test case selection unit that selects the test case to be executed in the regression test of the changed program,
The test case selection unit uses the dependency analysis result from the statement of the test case execution information, the expected value of the test case expected value information, and the difference statement, and the program before change Obtain a test case that may change the value of the expected value variable with the changed program, and select the obtained test case as the test case to be executed during the regression test.
A test case selection device characterized by that. The test case selector is
Select one test case from the test cases of the test case execution information,
Find the difference statement included in the statement executed in the selected test case,
Extracting the expected value variable of the selected test case from the test case expected value information,
Using the dependency analysis result, obtaining the statement affected by the obtained difference statement,
By obtaining a statement that changes the value of the extracted expected value variable from the obtained difference statement and the statement obtained using the dependency analysis result,
The test case selection device according to claim 1, wherein a test case that may change a value of the expected value variable is obtained. The test case selector is
Whether the statement obtained as a statement for changing the value of the extracted expected value variable is affected by the obtained difference statement when the selected test case is executed is used by using the dependency analysis result. Ask
When the statement obtained as a statement for changing the value of the extracted expected value variable is affected by the obtained difference statement, the selected test case is executed in the regression test. The test case selection device according to claim 2, which is selected as   The test case selection apparatus according to claim 1, wherein the program dependence analysis unit obtains a statement affected by the difference statement by examining at least an influence caused by a variable and an influence caused by a conditional branch.

Images