JPH10289102A - Data flow analyzer for program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To shorten the time needed for analysis of a data flow by extracting the information on definition/reference of a variable. SOLUTION: The functions of a syntax analysis means 26, a procedure decision means 21, a definition/reference information extraction means 23 and a data flow analysis means 24 are stored in a storage medium 31 as programs. These programs can be read and executed by a computer 30. The information on definition/reference of a variable is previously stored in a definition/reference information store file 22 by a prescribed procedure. This procedure includes a compiled and executable program, for example. If the prescribed information is acquired while the information on definition/reference of a variable is extracted for every line of a read source program, the said information is extracted by making reference to the file 22. Such a job is executed for the entire read source program.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data flow analyzer for a program used when debugging, testing, maintaining, etc., a source program described in a general programming language.

[0002]

2. Description of the Related Art Data flow analysis of a program is performed not only in debugging and testing after creating a source program but also in performing maintenance work such as modifying or remodeling an already created program as necessary. It is sometimes done. By performing data flow analysis when performing maintenance work, it is possible to efficiently extract necessary information, particularly when modifying a large-scale program. When performing such a data flow analysis, it is possible to perform a definition / reference analysis for determining a value definition location for each variable reference included in the source program and for determining a value reference location for variable substitution. is necessary.

[0003]

By the way, in many high-level languages such as the C language, various procedures are prepared in a standard library, and by specifying this procedure from a program, the procedure is executed. Can be called. It is also possible to call another procedure within one procedure. However, if another procedure is called in the procedure, data is passed there as well, so if you try to find out exactly the data dependencies, all the data dependencies inside the procedure called as the procedure will be changed. I have to find out. Such a task takes a long time to execute, especially in the case of a large-scale program.

When a procedure included in the standard library is provided as an executable program after being compiled, the contents of the source program do not exist and the procedure is created by using such a procedure. Data dependencies could not be determined for the program. Therefore, the conventional data flow analysis device cannot perform data flow analysis on such a program.

The present invention has been made based on the above circumstances, and enables data flow analysis even for a program including a procedure in which the contents of a source program cannot be viewed. It is an object of the present invention to provide a data flow analyzer for a program which can reduce the time required for data flow analysis even for a program including the program.

[0006]

In order to solve the above-mentioned problems, the present invention provides a definition / reference storing information on the definition / reference of a variable defined in a predetermined procedure described in a specific programming language. Information storage means, a procedure determining means for reading a source program written in the specific programming language, and determining where the predetermined procedure is included; and For a predetermined procedure, information relating to the definition / reference of a variable is read from the definition / reference information storage means, and for a procedure other than the predetermined procedure, information relating to the definition / reference of the variable is extracted from the contents of the source program. Reference information extracting means;
Data flow analysis means for performing a data flow analysis based on information relating to the definition / reference of the variable.

According to the present invention, as described above, information relating to the definition / reference of variables for a predetermined procedure is stored in the definition / reference information storage means in advance. In this case, the prescribed procedures include:
For example, an executable program after being compiled is also included. Then, when the information regarding the definition / reference of the variable is extracted line by line in the read source program, when the predetermined procedure is reached, the definition / reference information storage means is referred to for the definition / reference of the variable. Extract information. This operation is performed for the entire source program. Conventionally, even for a procedure that could not extract information on definition / reference of a variable without a source program, such information can be extracted, and subsequent data flow analysis can be performed. In this specification, in the case of "procedure",
It also includes functions in the C language.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows an example of a certain source program described in the C language. In the figure, the procedure “read” is described on the fourth line. This procedure is a separate procedure, "s
"canf" is called in the seventh line. here,
“Scanf” is a procedure included in the standard library of the C language, and “% d” specifies that the variable i to be defined is a decimal integer type.

Since the conventional data flow analyzer performs data flow analysis in a single procedure, one procedure contains a call statement for calling another procedure as shown in the program of FIG. In some cases, it is not possible to perform data flow analysis on a source program using a conventional device. When a procedure is called within a single procedure, it is common to perform “interprocedural analysis”, which results in a much faster execution speed than analysis within a single procedure. Slow down. Here, data flow analysis refers to analyzing which definition is referred to where.

In a normal data flow analysis within a single procedure, there is no data dependency in the procedure “read” on the fourth line, but “i” on the eighth line is actually
Data is dependent on i through the procedure “scanf” described in the seventh line. This is based on the specification of “scanf” included in the C language standard library, and the second argument i
Is defined. The “data dependence” means the following. When there is a statement S1, a statement S2, and a variable v, the variable v is defined in the statement S1. The variable v is referred to in the statement S2. An executable path from the statement S1 to the statement S2. The statement S2 is said to be data-dependent on the statement S1 with respect to the variable v if the three conditions that there is a path that does not redefine v exist.

The data flow analysis apparatus according to this embodiment is more applicable than the data flow analysis in a single procedure by utilizing the data flow analysis in a normal single procedure and the procedure information on a specific procedure. Conditions can be expanded, and the execution speed is faster than the interprocedural analysis, and accurate analysis can be performed. As described above, in the source program shown in FIG.
Since "f" is a procedure included in the standard library,
Its specifications are already known. That is, in the example of FIG. 1, it is known in advance that the variable i is defined. Therefore, for each procedure included in the standard library whose specification is clarified, such as “scanf”, the data dependency is checked in advance from the specification, and the definition / reference information describing only the result is described. Prepare a storage file. For example, in the case of “scanf”, the argument “i” stores information to be defined here in the definition / reference information storage file.

A procedure different from this, for example, "prin", which is a procedure for displaying the following argument on the screen, is described.
If the procedure is “tf”, this argument is “referenced” by this procedure. Therefore, "printf"
Is stored in the definition / reference information storage file.
As described above, in the definition / reference information storage file, for many procedures, a table is prepared in advance in which each procedure is associated with whether the variable is “definition” or “reference”. In this case, if a certain procedure is called, it is determined whether or not the procedure is included in this table. Extracts information on whether the called procedure is in this table and is a "definition" or a "reference" of a variable.

The procedures stored in the definition / reference information storage file are not limited to those included in the C language standard library. For example, if a particular factory creates a special procedure that is often used, and the procedure is generally used in the factory, the same procedure as the procedure included in the above standard library is used. If a definition / reference information storage file is created, an accurate and efficient analysis can be performed in the same manner.

Next, the operation of the program data flow analyzer of this embodiment will be described with reference to FIGS. FIG. 2 is a block diagram of the program data flow analyzer of the present embodiment, and FIG. 3 is a flowchart showing the operation. Note that the parsing means 2 shown in FIG.
6. Procedure determination means 21, definition / reference information extraction means 2
3. The data flow analysis means 24 can be realized on the computer 30 by reading the functions recorded in the storage medium 31 as programs on the storage medium 31. Therefore, this storage medium 31
Is assumed to be removable such as a flexible disk or a CD-ROM, by distributing the storage medium 31, the data flow analysis of the program of the present embodiment can be performed on any computer capable of reading and executing the storage medium 31. The device can be realized. At this time, the contents of the definition / reference information storage file 22 may be stored in the storage medium 31 together with these programs.

In the flowchart of FIG. 3, first, S1
At 0, the parsing means 26 generates a parse tree based on the input source program 20, and reads one sentence from the generated parse tree. At S11, the procedure determining means 2
1 judges whether or not this statement includes a specific procedure stored in the definition / reference information storage file 22;
If it is included, the definition / reference information extraction means 23 refers to the definition / reference information storage file 22 in S12,
In step S13, information as to whether the variable is “definition” or “reference” is extracted. On the other hand, when it is determined in S11 that the sentence does not include the procedure stored in the definition / reference information storage file 22, in S14, the definition / reference information extraction unit 23 determines whether the variable is “defined”. Determine if it is "referenced" or neither. In S15, the procedure determining means 21
Determines whether the statement of the program continues,
If it continues, the process returns to S10, where the information of the next sentence is read and the same operation as described above is executed. Further, when it is determined in S15 that the last sentence has been reached, in S16, the data flow analysis means 24 executes data flow analysis. The result is output as the analysis result 25.

As described above, even if the source program includes a procedure whose data dependency cannot be checked by the conventional data flow analyzer, necessary information is stored in the definition / reference information storage file 22 in advance. Write it down,
By referring to this as needed, it is possible to quickly examine data dependencies even for procedures that could not be examined before, or procedures that take a long time to execute when actually trying to examine them. Becomes

The present invention is not limited to the above embodiment, and various changes can be made within the scope of the invention. For example, in the above, “sc” included in the standard library of C language
Although typical procedures such as "anf" and "printf" have been described, the present invention can be applied to other procedures in the C language or procedures included in a library created by a user.
The same applies to many high-level languages other than languages.

[0018]

As described above, according to the present invention,
To enable data flow analysis for programs that include procedures that cannot see the contents of the source program, and to reduce the time required for data flow analysis even for programs that include procedures that allow you to view the source program. A data flow analysis device for a program capable of performing the above can be provided.

[Brief description of the drawings]

FIG. 1 is a diagram showing an example of a certain source program described in C language.

FIG. 2 is a block diagram of a program data flow analyzer according to an embodiment of the present invention.

FIG. 3 is a flowchart showing the operation of the program data flow analysis device of one embodiment of the present invention.

[Explanation of symbols]

 Reference Signs List 20 source program 21 procedure determination means 22 definition / reference information storage file 23 definition / reference information extraction means 24 data flow analysis means 25 analysis result 26 syntax analysis means 30 computer 31 storage medium

Claims (3)

[Claims] 1. A definition / reference information storing means for storing information relating to definition / reference of a variable defined in a predetermined procedure described in a specific programming language, and a source program described in the specific programming language. A procedure judging means for reading and judging whether the predetermined procedure is included, and defining / referencing a variable from the definition / reference information storage means for the predetermined procedure based on a judgment result of the procedure judging means. Definition / reference information extracting means for reading information on the definition / reference of a variable from the contents of the source program for procedures other than the predetermined procedure, and data based on the information on the definition / reference of the variable. Data flow analysis means for performing flow analysis; and a program comprising: Data flow analysis apparatus. 2. The data flow analysis device according to claim 1, wherein the predetermined procedure includes an executable program. 3. A computer, comprising: a definition / reference information storage means for storing information relating to definition / reference of variables defined in a predetermined procedure described in a specific programming language; A procedure determining means for reading a source program and determining whether the predetermined procedure is included; and, based on a determination result of the procedure determining means, the variable of the predetermined procedure from the definition / reference information storage means. Definition / reference information extracting means for reading information relating to definition / reference and extracting information relating to definition / reference of variables from the contents of the source program for procedures other than the predetermined procedure; To function as a data flow analysis means that performs data flow analysis based on Recorded machine-readable storage medium a program.