JP4724387B2 - Program conversion program, program conversion apparatus, and program conversion method - Google Patents

Description

  The present invention relates to a program conversion program, a program conversion apparatus, and a program conversion method for converting a starting source program described in a first programming language into a target source program described in a second programming language, and particularly has high maintainability. The present invention relates to a program conversion program, a program conversion apparatus, and a program conversion method capable of obtaining a target source program.

  In recent years, due to problems such as high operational costs and low development efficiency, it has been widely used to replace a system that operates on a host computer with a so-called open system that operates on a UNIX (registered trademark) server or a PC server. Yes. In replacement, a new system can be re-developed, but a conventional system is often used to save the time and cost required for migration.

  However, even when the system is used, the problem of low development efficiency cannot be solved by simply compiling the programming used on the host computer in a new environment. Therefore, at the time of migration, it is necessary to convert a program written in a language for a host computer such as COBOL or PL / I into a more productive language such as C ++ or Java (registered trademark).

  Conventionally, a technique for converting a program written in one programming language into a program written in another programming language is known. For example, Patent Document 1 discloses a technique for converting a program using a conversion rule. Patent Document 2 discloses a technique for making conversion rules more versatile and realizing conversion of more programming languages.

Japanese Patent Laid-Open No. 02-010434 Japanese Patent Publication No. 07-078742

  However, since these conventional techniques simply perform lexical conversion so that the processing contents of the program language are equal, there is a problem that the converted program has low maintainability.

  Each programming language has its own characteristics, and it is usually programmed in a style that matches those characteristics. However, a program that is simply lexically converted inherits the style of the programming language before conversion as it is, and the description becomes unnatural as a programming language after conversion. Such a program is extremely low in readability for a developer who does not know the language before conversion, and is not desirable from the viewpoint of maintainability.

  In particular, an unnatural source program is generated when converting between languages with very different linguistic ideas, such as when converting from procedural language COBOL to object-oriented language Java (registered trademark). Very likely. Programs usually require additional functions or modifications after migration, and poor maintainability is a major problem in program maintenance.

  The present invention has been made to solve the above-described problems caused by the prior art, and provides a program conversion program, a program conversion apparatus, and a program conversion method capable of efficiently obtaining a highly maintainable program as a conversion result. The purpose is to do.

In order to solve the above-described problems and achieve the object, a program conversion program according to the present invention converts an origin source program described in a first programming language into a target source program described in a second programming language. A pre-conversion model generation procedure that parses the starting source program and generates a syntax tree of the starting source program based on the grammar of the first programming language , and stores it in a first storage unit been, the first based on a syntax tree based on the grammar of the programming language conversion rule for converting the syntax tree based on the grammar of the second programming language, the syntax tree of the origin source program second programming language in converting the syntax tree based on the grammar, stored in the second storage unit Based hierarchical data type condition to exclude the conversion based on the conversion rule corresponding to the syntax tree of the origin source program including a portion corresponding to the condition, grammar of the second programming language A post-conversion model generation procedure for inserting clue information indicating that it is excluded from conversion based on the conversion rule, and a syntax tree based on the grammar of the second programming language. in converting into the desired source program written in a language, embedding the key information, the object source program, the location where the cue information corresponding to syntax tree based on the grammar of the inserted second programming language A source program generating procedure is executed by a computer.

The program conversion apparatus according to the present invention is a program conversion apparatus for converting a starting source program described in a first programming language into a target source program described in a second programming language, wherein the starting source program And a pre-conversion model generation means for generating a syntax tree of the source program based on the grammar of the first programming language, and a syntax tree based on the grammar of the first programming language. A first storage unit that stores a conversion rule for converting to a syntax tree based on the grammar of the second, a second storage unit that stores a condition for excluding a hierarchical data type from conversion based on the conversion rule, based on the conversion rule stored in the first storage unit, the syntax tree of the second of said source program Upon conversion to a syntax tree based on the grammar of the programming language, corresponding to the syntax tree of the origin source program including a portion corresponding to the conditions stored in the second storage section, the syntax of the second programming language the syntax tree based on the excluded from conversion based on the conversion rule and the converted model generating means for inserting a key information indicating that the said second of said second programming data model based on the grammar of the programming language in converting into the desired source program written in a language, embedding the key information, the object source program, the location where the cue information corresponding to syntax tree based on the grammar of the inserted second programming language And a source program generation means.

The program conversion method according to the present invention is a program conversion method for converting a starting source program described in a first programming language into a target source program described in a second programming language, the starting source program a pre-conversion model generation step of parsing to generate a parse tree of the origin source program based on the grammar of the first programming language, stored in the first storage section, the syntax of the first programming language in based the syntax tree conversion rule for converting the syntax tree based on the grammar of the second programming language, converts the syntax tree of the origin source program syntax tree based on the grammar of the second programming language based on, The hierarchical data type stored in the second storage unit is based on the conversion rule. Under the terms of interest out of the conversion corresponds to the syntax tree of the origin source program including a portion corresponding to the condition, the syntax tree based on the grammar of the second programming language, the conversion based on the conversion rule A post-conversion model generation step for inserting clue information indicating that it is excluded, and a syntax tree based on the grammar of the second programming language is converted into the target source program described in the second programming language And a source program generation step of embedding the clue information in an appropriate location of the target source program.

  According to the present invention, if a program conversion is performed by simple lexical conversion, a confirmation required tag is output to a corresponding portion for a portion that is likely to generate a complicated and low-readable code. Since the program conversion is performed by using a tag as a clue in a different method from simple lexical conversion, a highly maintainable conversion result can be obtained.

  In the program conversion program according to the present invention as set forth in the invention described above, the source program generation procedure embeds the clue information in the target source program as a program comment.

  According to the present invention, clue information is output as a program comment at a place where an appropriate code needs to be written, and the program can be edited and compiled as usual while the clue information exists. Because it is configured, program conversion can be performed efficiently.

  The program conversion program according to the present invention is the program conversion program according to the above invention, wherein the source program generation procedure includes the name of the starting source program as the clue information and the position of the starting source program corresponding to the predetermined rule. It is characterized by embedding information.

  According to the present invention, since the conversion source source can be easily referred to using the clue information, the program can be converted efficiently.

  According to the present invention, if a program conversion is performed by simple lexical conversion, a confirmation required tag is output to a corresponding portion for a portion that is likely to generate a complicated and low-readable code. Since the program conversion is performed by a method different from simple lexical conversion using the tag as a clue, it is possible to obtain a conversion result with high maintainability.

  Further, according to the present invention, since the conversion source source can be easily referred to using the clue information, the program can be converted efficiently.

  Exemplary embodiments of a program conversion program, a program conversion apparatus, and a program conversion method according to the present invention will be explained below in detail with reference to the accompanying drawings. Here, a case where a program written in COBOL is converted into a program written in Java (registered trademark) will be described as an example. However, the present invention does not depend on a specific language, and various It can be used for program language conversion.

  First, the outline of the present invention will be described in comparison with a conventional method. FIG. 14 is an explanatory diagram for explaining a conventional programming conversion method. As shown in the figure, in the conventional conversion method, the input unit first reads a source program before conversion and performs syntax analysis, and a syntax tree of the source program before conversion (hereinafter referred to as “source data model before conversion”). Create

  Subsequently, the conversion engine converts the pre-conversion source data model based on a predetermined conversion rule to create a syntax tree of the post-conversion source program (hereinafter referred to as “post-conversion data source model”). Then, the output unit generates a converted source program from the converted data source model, and the conversion process is completed.

  Here, since the conversion rules used by the conversion engine are mechanical lexical conversions so that the processing contents of the program are the same, the source program after conversion is affected by the language specifications of the programming language before conversion. May result in unnatural and poorly readable code. In order to make such a program highly maintainable, it is necessary for a developer to analyze the generated code over time and replace a complicated part with a natural code.

  FIG. 1 is an explanatory diagram for explaining a conversion method according to the present embodiment. As shown in the figure, the conversion method according to the present embodiment is different from the conventional method in that a program is converted using a confirmation required rule in addition to the conversion rule. The confirmation-needed rule is a rule that, when a predetermined condition is met, another correspondence is performed without performing mechanical lexical conversion by the conversion rule.

  In this method, the pre-conversion source data model created by the input unit is converted into a post-conversion source data model by the conversion engine based on the conversion rule and the confirmation required rule. In the converted source data model, the confirmation required part that matches the confirmation required rule is replaced with a confirmation required tag. The confirmation required tag is embedded in an appropriate position of the converted source program in the form of a comment or the like by the output unit.

  The required confirmation tag embedded in the converted source program is replaced with an appropriate description in the required-conversion-point conversion unit to become a source program that can be actually operated. Since the replacement of the description in the confirmation point conversion unit may require business knowledge, it is usually performed manually, but conversion based on the business knowledge can be ruled and automatically converted.

  By adopting such a method, information indicating what correction is necessary is embedded in the part that has been converted into a difficult code in the conventional method, and the developer can use this as a clue to find the appropriate source. The code can be completed efficiently, and as a result, a highly maintainable program can be obtained at a lower cost than before.

  Next, an example of program conversion by the program conversion method according to the present embodiment is shown. FIG. 2 is an explanatory diagram for illustrating an example of program conversion by the program conversion method according to the present embodiment.

  There are many differences between COBOL and Java (registered trademark) in terms of language specifications, but there are also differences in basic storage area management methods. In COBOL, one storage area can be divided hierarchically and given a name for each hierarchy. For example, in the COBOL source program before conversion in FIG. 2, the area named WK-DATE is divided into an area named YY, an area named MM, and an area named DD. Yes. If necessary, the area named YY can be further divided into an area named YY1 and an area named YY2.

  This hierarchical storage area (hereinafter referred to as “group item”) shares the same storage area. For example, if the value of the area named MM is updated, one of the areas named WK-DATE is updated. The department has been updated. Also, if the entire area named WK-DATE is updated, the area named YY, the area named MM, and the area named DD are updated respectively. A program created by COBOL using such a language specification refers to and updates storage areas using various hierarchies as needed.

  However, in Java (registered trademark), it is impossible to divide a single storage area hierarchically in terms of language specifications, and therefore, how to perform conversion becomes a problem. It is possible to prepare storage areas for each layer separately and devise access methods to make them appear to be shared, but the source program becomes very complex.

  Therefore, in the example of FIG. 2, simple lexical conversion is not performed in the group item reference and update processing according to the conversion rule required. For example, the COBOL command statement “MOVE DATE TO WK-DATE.” Indicates a process of transferring the value of a group item to another group item. This process is a Java (registration) generated by the output unit. (Trademark) source program is converted as an empty set access method, and a comment indicating the processing content to be implemented is inserted therein.

  This empty access method is implemented with necessary processing by a developer who has business knowledge, and only setting processing for areas required for business is added, so a program with high maintainability can be obtained without complicating the program. Will be.

  Next, the configuration of the program conversion apparatus 100 according to the present embodiment will be described. FIG. 3 is a functional block diagram illustrating the configuration of the program conversion apparatus according to the present embodiment. As shown in the figure, the program conversion apparatus 100 according to the present embodiment includes a user interface unit 200, a control unit 300, and a storage unit 400.

  The user interface unit 200 is a device that provides information to the user and receives instructions from the user, and includes a display unit 210 including a display device such as a liquid crystal device, a keyboard, a mouse, and the like. And an operation unit 220.

  The control unit 300 is a control unit that totally controls the program conversion apparatus 100, and includes a pre-conversion cleansing unit 310, a pre-conversion model generation unit 320, a post-conversion model generation unit 330, a source program generation unit 340, and confirmation required. The location conversion unit 350, the rule extraction unit 370, and the work estimation unit 380 are included.

  The pre-conversion cleansing unit 310 is a processing unit that replaces the pre-conversion source program with an equivalent expression that is easier to convert. The source program before conversion often has a complicated process flow through several modifications, and if this is converted as it is, the process becomes difficult to understand. Therefore, the process flow of the source program is arranged and corrected so that it becomes a simple flow before conversion.

  FIG. 4 is an explanatory diagram illustrating an example of pre-conversion cleansing. As shown in the figure, in the source program 10 before the conversion by the pre-conversion cleansing unit 310, the processing execution order is the order of processing 1 → processing 2 → processing 3. Process 1 → Process 3 → Process 2 in this order, and the process description order and the actual process flow are different. Accordingly, the pre-conversion cleansing unit 310 changes the process description order from process 1 → process 3 → process 2 so that the process description order matches the actual process flow, and the converted program is easy to understand. It is trying to become.

  The pre-conversion model generation unit 320 is a processing unit that parses a pre-conversion source program and generates a pre-conversion source data model, and corresponds to the input unit of FIG. The pre-conversion source data model is a model that logically represents the structure of the pre-conversion source program, and is used as an intermediate format for language conversion.

  In the present embodiment, the source data model before and after conversion is expressed using XML (Extensible Markup Language). XML is a format that can express data and its structure at the same time using a marker character string called a tag.

  FIG. 5 is an explanatory diagram showing an example of conversion from a pre-conversion source program to a pre-conversion source data model. As shown in the figure, the pre-conversion model generation unit 320 creates a program tag as the pre-conversion source data model 40, sets “PG1” as the program name of the pre-conversion source program 30 as its name attribute, and field. “1” which is the ID of this model is set as an attribute.

  Then, a data_division tag 41 is created as an element of the program tag, and the analysis result of the data part 31 of the source program 30 before conversion is stored therein. Since the variable A definition 31a defines the content defined in the source program A as A, the content of the source program A is also referred to and stored. Specifically, the data_definition tag 41a added as an element of the data_division tag 41 is a variable identified by an ID having a variable A of “1”, and the first line of a program file having a field of “100”. Indicates a variable of level 1 defined in FIG.

  In this example, the variable A is a COPY clause, and the definition is in a separate file, so this is the expression. If the variable A is defined in the same file as “01 A PIC X (2).”, The data_definition tag 41a becomes an expression indicating that the variable A is a two-digit character string.

  Further, the pre-conversion model generation unit 320 creates a procedure_division tag 42 as an element of the program tag, and stores the analysis result of the program unit 32 of the pre-conversion source program 30 therein.

  In this example, the MOVE sentence 32 a is analyzed, and the analysis result is stored as a statement tag 42 a that is an element of the procedure_division tag 42. Specifically, in the statement tag 42a, the type of the instruction statement is a MOVE statement, the description position in the source program 30 before change is the 20th line, and the transfer source of the MOVE statement is identified by the ID “1”. The variable A indicates that the transfer destination is the variable B identified by the ID “2”.

  When converting a program, it is necessary to examine the usage status of variables over the entire program. Therefore, it is necessary to convert related programs at once, but the number of programs to be converted It is not uncommon to range from hundreds to thousands. For this reason, even when it is necessary to add one program to be converted, a large amount of processing time is required to parse a large amount of programs again and make a source data model. Become.

  Therefore, the pre-conversion model generation unit 320 stores the generated pre-conversion source data model in the source data model storage unit 410, and stores it here unless a new or updated source program is designated as a conversion target. Read and output the converted source data model.

  The post-conversion model generation unit 330 is a processing unit that converts the pre-conversion source data model based on information in the rule storage unit 420 and generates a post-conversion source data model, and corresponds to the conversion engine of FIG. The source data model before conversion is a model based on the language specification of COBOL, whereas the source data model after conversion is a model based on the language specification of Java (registered trademark).

  The basic processing content of the post-conversion model generation unit 330 is based on the information in the conversion rule DB 423, replacing the part of the pre-conversion source data model that matches the information in the confirmation rule DB 421 with a confirmation tag. Conversion from a COBOL model to a Java (registered trademark) model. There are several problems in converting from a COBOL model to a Java (registered trademark) model.

  The first problem is a data type problem. For example, the COBOL character string type has the concept of the number of digits. If the variable X is a four-digit character string type variable and the variable Y is a two-digit character string type variable, the same “ABC” is used for both. When a comparison is made by substituting values, it is determined that the variables X and Y do not match due to the difference in the number of digits.

  On the other hand, since there is no concept of the number of digits in the String class that is a standard character string type in Java (registered trademark), when the same value “ABC” is substituted into the variables X and Y of the String class, Both are determined to match.

  In order to solve such a data type problem, it is possible to replace the comparison processing at the time of conversion so that the operation is the same, but the processing after conversion may be complicated, and maintainability Unfavorable above. Therefore, in the program conversion according to the present embodiment, in order to solve such a data type problem, a wrapper type that behaves like the data type after conversion and reflects the characteristics of the data type before conversion is set to Java. Prepared on the (registered trademark) side and converted the program using this wrapper type. If such a wrapper type has the function of the language before conversion, equivalent processing can be realized by simple conversion by simply replacing the type before conversion with the wrapper type, and a simple and highly maintainable program can be obtained.

  Another problem is how to access variables. In object orientation, direct access to variables in a class from the outside is considered undesirable from the viewpoint of data hiding, which is the basic concept of object orientation, and usually a method of accessing variables via methods. Is taken. However, since there is no such convention in COBOL, the conventional simple conversion method generates a program that does not look like Java (registered trademark) that directly accesses variables, or accesses all variables. The method of generating the method was taken.

  Therefore, in the program conversion according to the present embodiment, the reference / update relationship of variables is derived from the source data model before conversion, only the access methods required at the time of conversion are generated, and direct access to the variables is accessed. Replaced with access via method. Thereby, since only necessary and sufficient access methods are created, it is possible to obtain a Java (registered trademark) program having higher maintainability than the conventional method.

  FIG. 6 is an explanatory diagram illustrating an example of conversion from a source data model before conversion to a source data model after conversion. As shown in the figure, the post-conversion model generation unit 330 creates a class tag as the post-conversion source data model 50, and the name attribute and field attribute of the program tag of the pre-conversion source data model 40 in its name attribute and classid attribute. Set the value of.

  Then, a field tag 53 is created as an element of the class tag, and the result of converting the contents of the data_division tag 41 of the source data model 40 before conversion using information in the rule storage unit 420 is stored therein.

  The data_definition tag 41a indicates that the type of the variable is defined in an external file. Therefore, the data_definition tag 41a is converted as a variable of a class defined in another file, and the information is converted into a field_definition tag 53a that is an element of the field tag 53. Store as. The field_definition tag 53a is specifically a variable identified by the variable a having an ID of “1”, and the class of this variable indicates that the classid is “100”.

  Here, the reason for converting the variable name from uppercase to lowercase is to match that the variable is normally written in lowercase in Java (registered trademark), and a program that is easy to read for general Java (registered trademark) programmers. It is a process for generating.

  Further, the post-conversion model generation unit 330 creates a method tag 54 as an element of the class tag, and converts the contents of the procedure_division tag 42 of the pre-conversion source data model 40 using the information in the rule storage unit 420. Store.

  In this example, the variable content transfer process is converted and stored as a statement tag 54 a that is an element of the method tag 54. The statement tag 42a means that the contents of the variable are transferred by assignment. However, as described above, in Java (registered trademark), there is a convention that a variable is not directly accessed between classes. It is replaced with the process of posting the variable contents by calling. Specifically, the processing is replaced with a process of calling the set method of the transfer destination variable b and passing the transfer source variable a as an argument of the method.

  The entity of the set method is inserted into the converted source data model 51 of class B. The post-conversion source data model 51 of class B is a model generated by converting the source program incorporated in the COPY phrase in the pre-conversion source program 30, and initially includes only variable type definitions. . In an object-oriented language such as Java (registered trademark), the definition of a variable and the process for operating the defined variable need to be contained in one class. Let's do it.

  Since the set method is a method for accessing a variable, it is added as a field_access tag 56 that is an element of the field tag 55. Here, the process of the conversion source statement tag 42a is a process of transcribing a value to the group item, and corresponds to one of the rules of the confirmation required rule DB 421. For this reason, the processing content of the statement tag 42a is not converted based on the conversion rule DB 423, but an empty set method 56a is created, and a required check tag 56b is inserted into the empty set method 56a based on the required check rule DB 421.

  For the confirmation tag 56b, specifically, the process to be embedded in this place is described in the 20th line of the conversion source program “PG1.cob”, and “TODO-1” is used to assist the developer who converts this part. This means that it is necessary to display the message indicated by the comment ID “”.

  In this way, the post-conversion model generation unit 330 creates only access methods that are actually necessary. Since an access method that is not used is not created, an unnecessary access method does not exist in the converted program, and high maintainability can be obtained.

  The source variable of the field_definition tag 55a of the post-conversion source data model 51 of class B is a character type. As already described, the COBOL character string type has the concept of the number of digits, and in the Java (registered trademark) program after conversion, it is necessary to convert it into a wrapper class of the String class having the concept of the number of digits. . For this reason, the number of digits of the conversion source is held in the field_definition tag 55a as a size attribute.

  The source program generation unit 340 is a processing unit that generates a post-conversion source program with a confirmation part from the post-conversion source data model, and corresponds to the output unit of FIG. The post-conversion source program with the confirmation required part is a source program that outputs the confirmation necessary tag of the converted source data model as a comment.

  FIG. 7 is an explanatory diagram showing an example of conversion from the converted source data model to the converted source program. As shown in the figure, the source program generation unit 340 converts the field_definition tag 53a of the converted source data model 50 into a variable definition 60a, converts the statement tag 54a into a set statement 60c, and so on. Is generated.

  The converted source data model 51 is converted to generate a converted source program 61. Since the post-conversion source data model 51 includes a confirmation tag 56b, the comment corresponding to the comment ID set in the comment attribute is acquired with reference to the confirmation comment DB 422, and this comment and the process to be embedded here Information indicating where in the pre-conversion source program exists as a comment at a predetermined position. The field_definition tag 55a that defines a character type variable is converted not as a String class variable but as a BizString class variable having the concept of the number of digits.

  The required part conversion unit 350 is a processing unit that replaces the required confirmation part embedded in the form of a comment in the converted source program with the required part with an appropriate description and completes the converted source program. In the present embodiment, the confirmation required part conversion unit 350 is implemented as an editing screen for the developer to edit the converted source program with the confirmation necessary part. The developer performs editing with reference to the guide shown in FIG.

  The type conversion unit 360 is a processing unit that replaces the wrapper class included in the converted source program with a general class. As described above, in order to absorb the difference between the data types of COBOL and Java (registered trademark), the program conversion according to the present embodiment uses the Java (registered trademark) model that imitates the data type of COBOL. May be replaced with a wrapper class.

  However, for example, in the case of the COBOL character string type, if there is no comparison between strings or implicit digit reduction, it is converted to the String class, which is the most common class that handles character strings instead of the wrapper class. No problem. The wrapper class does not greatly impair the maintainability of the converted program, but the readability is enhanced by using a more universal class.

  Therefore, the type conversion unit 360 refers to the model stored in the source data model storage unit 410, and converts a variable that is determined to be a normal class variable instead of a wrapper class to a normal class variable. To do. FIG. 9 is an explanatory diagram illustrating an example of type conversion by the type conversion unit. As shown in the figure, the variable definition 64a using the BizString class is determined not to be used for comparison, and is converted into a variable definition 71a using a general String class.

  The rule extraction unit 370 is a processing unit for extracting a pattern that frequently appears from the editing contents performed by the confirmation point conversion unit 350 and feeding it back to the conversion rule DB 423 to automate the subsequent conversion process and improve efficiency. is there. The rule extraction unit 370 measures, for example, the degree of similarity of the process embedded in the confirmation required portion, and extracts a similar pattern that appears frequently.

  FIG. 10 is an explanatory diagram showing an example of program conversion based on the fed back rules. This figure shows an example in which a rule for converting date-related data is fed back. In COBOL, the date is often defined as a variable consisting of an 8-digit number with a name such as YYMMDD, and the storage area of this variable is further divided to define variables with names such as YY, MM, and DD. , Used as year, month and date respectively.

  Such date-related group items exist in various programs, and if they are manually replaced one by one, the work cost will be high. By preparing a general-purpose class such as BizDate and setting the confirmation rule DB 421 and the conversion rule DB 423 so as to automatically perform program conversion using this class, the program conversion work can be made more efficient.

  The work estimation unit 380 is a processing unit that estimates the time required for conversion by the confirmation required unit. If the source program with required confirmation points output by the source program generation unit 340 contains a large number of necessary confirmation points, the work time required for program conversion is the same as when the program is re-developed from scratch. It may become.

  In such a case, it is wise to choose development from scratch that can provide a more maintainable program. Therefore, the work estimation unit 380 uses the information in the source data model storage unit 410 and the confirmation rule DB 421 to estimate how many confirmation points are generated and how much to determine how to deal with program conversion. To do.

  The storage unit 400 includes a source data model storage unit 410 and a rule storage unit 420. The source data model storage unit 410 stores the pre-conversion data source model generated by the pre-conversion model generation unit 320 and the post-conversion data source model generated by the post-conversion model generation unit 330.

  The rule storage unit 420 includes a confirmation rule DB 421, a confirmation comment DB 422, a conversion rule DB 423, and a type library DB 424. The confirmation required rule DB 421 stores a rule for a case where lexical conversion based on the conversion rule DB 423 is not performed.

  FIG. 11 is a data structure diagram illustrating an example of the data structure of the confirmation rule DB 421 that is required. As shown in the figure, the confirmation required rule DB has items of a statement, a condition, a corresponding location, and a comment ID. The statement stores an imperative sentence targeted by the rule. The condition stores a condition for applying the rule. The corresponding location stores the insertion destination of the confirmation-required tag when the rule is applied. The comment ID stores a comment ID that is embedded in a confirmation required tag when the rule is applied.

  The required comment DB 422 stores a comment corresponding to the comment ID of the required rule DB 421. FIG. 12 is a data structure diagram showing an example of the data structure of the confirmation comment DB 422 required. As shown in the figure, the required comment DB 422 has items of comment ID and comment. The comment ID stores a value corresponding to the comment ID of the confirmation required rule DB 421. The comment stores a comment sentence that is embedded in the source program with a confirmation part when the confirmation rule is applied. Note that the #filename and #line in the comment text are replaced by the source program generation unit 340 with the program name before conversion and the line number of the conversion part, respectively.

  The conversion rule DB 423 stores rules for lexically converting the pre-conversion source data model. The type library DB 424 stores wrapper class information used at the time of program conversion in pairs with the data type of the conversion source.

  Next, the processing procedure of the program conversion apparatus 100 shown in FIG. 3 will be described. FIG. 13 is a flowchart showing a processing procedure of the program conversion apparatus 100 shown in FIG.

  As shown in the figure, when the source program before conversion is read (step S101), it is confirmed whether or not the program is new or updated. If the program is new or updated (Yes at Step S102), a cleansing process is performed to correct the program so that the conversion is easy (Step S103). If the program is neither new nor updated (No at Step S102), Step S103 is not performed.

  If there is a file that has not been read yet (No at Step S104), the process returns to Step S101 to read the next source program. If all the source programs before conversion have been read (Yes at Step S104), a source data model before conversion is generated (Step S105), the entire source data model before conversion is analyzed, and the converted source data is converted. A model is created (step S106). In this post-conversion source data model, a confirmation tag is inserted in a portion that is likely to be converted as a difficult code by lexical conversion.

  Then, a source program with a confirmation required part is generated based on the converted source data model (step S107). In this source program with a confirmation part, the contents of the confirmation tag are inserted in the form of a comment.

  Then, an appropriate code is embedded in the part of the confirmation tag of the source program with confirmation part required (step S108), the non-necessary wrapper class is converted into a normal class, and the conversion of the program is completed (step S109).

  As described above, in the present embodiment, when the post-conversion model generation unit 330 generates the post-conversion source data model from the pre-conversion source data model, the part corresponding to the rule of the confirmation required rule DB 421 is the conversion rule DB. The lexical conversion used is not performed, and auxiliary information for embedding appropriate code is embedded, so if simple lexical conversion is performed, processing that is likely to generate complicated and low-readable code is acquired. It is possible to efficiently convert with high-quality code as a background, and thus it is possible to efficiently obtain a post-conversion source program with high maintainability.

(Supplementary Note 1) A program conversion program for converting a source program written in a first programming language into a target source program written in a second programming language,
A pre-conversion model generation procedure for parsing the origin source program to generate a data model based on the first programming language;
Location corresponding to a predetermined rule for determining that the converted data model may be complicated when the data model based on the first programming language is converted to the data model based on the second programming language In addition, a post-conversion model generation procedure for inserting clue information for performing conversion by another means,
When converting a data model based on the second programming language into the target source program described in the second programming language, a source program generation procedure for embedding the clue information in an appropriate portion of the target source program; A program conversion program which is executed by a computer.

(Supplementary note 2) The program conversion program according to supplementary note 1, wherein the source program generation procedure embeds the clue information in the target source program as a comment of the program.

(Additional remark 3) The source program generation procedure embeds the name of the starting source program and the position information of the portion corresponding to the predetermined rule in the starting source program as the clue information. 2. The program conversion program according to 2.

(Supplementary Note 4) The converted model generation procedure is performed by the second programming language when the data type corresponding to the data type used in the starting source program does not exist in the second programming language. 4. The program conversion program according to appendix 1, 2, or 3, wherein the conversion is performed using a specified data type created in advance so as to have a specification equivalent to the type.

(Supplementary Note 5) If it is determined that there is no problem in replacing the specified data type with a general-purpose data type that is normally used, type conversion that replaces the former data type included in the target source program with the latter data type The program conversion program according to appendix 4, wherein the program is further executed by a computer.

(Supplementary note 6) In the preceding stage of the pre-conversion model generation procedure, the computer further executes a pre-conversion cleansing procedure for equivalently transforming the starting source program into a description that is easier to convert. The program conversion program as described in any one.

(Supplementary note 7) The post-conversion model generation procedure creates a method for accessing a variable on the basis that there is an actual access to the variable when the second programming language is an object-oriented language. The program conversion program according to any one of supplementary notes 1 to 6, characterized in that:

(Supplementary Note 8) A computer further executes a storage procedure for storing a data model based on the first programming language generated by the pre-conversion model generation procedure,
In the pre-conversion model generation procedure, when the same starting source program as the one converted in the past is accepted as a conversion target, the conversion source data model stored in the previous storage procedure is stored without performing the conversion process. The program conversion program according to any one of appendices 1 to 7, wherein the program conversion program is characterized in that

(Additional remark 9) The work estimation procedure which totals the number of the said clue information embedded in the said target source program according to the kind of clue information based on the data model based on the said 1st programming language produced | generated by the said pre-conversion model production | generation procedure. The program conversion program according to any one of appendices 1 to 8, further causing the computer to execute the program.

(Supplementary note 10) Any one of Supplementary notes 1 to 9, further causing a computer to execute a rule extraction procedure for extracting a conversion rule from a description of a process added to the clue information portion of the target source program Program conversion program described in 1.

(Supplementary note 11) A program conversion device for converting a source program written in a first programming language into a target source program written in a second programming language,
Pre-conversion model generation means for parsing the starting source program to generate a data model based on the first programming language;
Location corresponding to a predetermined rule for determining that the converted data model may be complicated when the data model based on the first programming language is converted to the data model based on the second programming language In addition, after-conversion model generation means for inserting clue information for performing conversion by another means,
Source program generation means for embedding the clue information in an appropriate portion of the target source program when converting a data model based on the second programming language into the target source program described in the second programming language; A program conversion apparatus comprising:

(Supplementary note 12) A program conversion method for converting an origin source program described in a first programming language into a target source program described in a second programming language,
A pre-conversion model generation step of parsing the origin source program to generate a data model based on the first programming language;
Location corresponding to a predetermined rule for determining that the converted data model may be complicated when the data model based on the first programming language is converted to the data model based on the second programming language In addition, a post-conversion model generation process for inserting clue information for performing the conversion in a separate process,
A source program generation step of embedding the clue information in an appropriate portion of the target source program when converting a data model based on the second programming language into the target source program described in the second programming language; The program conversion method characterized by including.

  As described above, the program conversion program, the program conversion apparatus, and the program conversion method according to the present invention are useful for program language conversion, and are particularly suitable for program conversion that requires a highly maintainable source program as a conversion result. Yes.

It is explanatory drawing for demonstrating the conversion system which concerns on a present Example. It is explanatory drawing for showing an example of the program conversion by the program conversion system based on a present Example. It is a functional block diagram which shows the structure of the program conversion apparatus which concerns on a present Example. It is explanatory drawing which shows an example of the cleansing before conversion. It is explanatory drawing which shows an example of the conversion from the source program before conversion to the source data model before conversion. It is explanatory drawing which shows an example of the conversion from the source data model before conversion to the source data model after conversion. It is explanatory drawing which shows an example of the conversion from the source data model after conversion to the source program after conversion. It is an example of a confirmation tag conversion rule guide required. It is explanatory drawing which shows an example of the type conversion by a type conversion part. It is explanatory drawing which shows an example of the program conversion by the rule which fed back. It is a data structure figure which shows an example of the data structure of confirmation required DB. It is a data structure figure which shows an example of the data structure of confirmation required DB. It is a flowchart which shows the process sequence of the program conversion apparatus shown in FIG. It is explanatory drawing for demonstrating the conventional programming conversion system.

Explanation of symbols

10 source program 20 source program after cleansing 30 source program before conversion 31 data part 31a definition of variable A 31b definition of variable B 32 program part 32a MOVE sentence 40 source data model before conversion 41 data_definition tag 41b data_definition tag 41b data_definition tag 41b data_definition tag 42a statement tag 50 source data model after conversion 51 source data model after conversion 52 source data model after conversion 53 field tag 53a field_definition tag 53b field_definition tag 54 method tag 54a statement tag 55 field tag 55 eld_definition tag 55b field_definition tag 56 field_access tag 56a empty set method 56b confirmation required tag 60 converted source program 60a variable definition 60b variable definition 60c set statement 61 converted source program 61a variable definition 61c post conversion method 62c variable definition 61c access method 70 Source program after conversion 70a Variable definition 70b Variable definition 70c set statement 71 Source program after conversion 71a Variable definition 71b Variable definition 71c Access method 72 Source program after conversion 80 Source program before conversion 80a Definition of date 80b Operation of date 81 PG1 class without feedback 81a Definition of year / month / day 81b Operation of year / month / day 8 2 Date class created individually 83 Class PG1 with feedback 83a Definition of year, month, day 83b Operation of year, month, day 84 Shared date class 100 Program converter 200 User interface unit 210 Display unit 220 Operation unit 300 Control unit 310 Cleansing unit before conversion 320 Model generation unit before conversion (input unit)
330 Model generator after conversion (conversion engine)
340 Source program generator (output unit)
350 Confirmation point conversion unit 360 Type conversion unit 370 Rule extraction unit 380 Work estimation unit 400 Storage unit 410 Source data model storage unit 420 Rule storage unit 421 Required rule DB
422 Comment DB required
423 Conversion rule DB
424 type library DB

Claims (6)

A program conversion program for converting an origin source program described in a first programming language into a target source program described in a second programming language,
A pre-conversion model generation procedure for parsing the starting source program to generate a syntax tree of the starting source program based on the grammar of the first programming language;
Stored in the first storage unit, based on a syntax tree based on the grammar of the first programming language conversion rule for converting the syntax tree based on the grammar of the second programming language, the syntax tree of the origin source program Is converted into a syntax tree based on the grammar of the second programming language, based on the condition that the layered data type stored in the second storage unit is excluded from the conversion based on the conversion rule, The clue information indicating that the syntax tree based on the grammar of the second programming language corresponding to the syntax tree of the starting source program including the portion corresponding to the condition is excluded from the conversion based on the conversion rule. The post-conversion model generation procedure to be inserted,
When converting a syntax tree based on the grammar of the second programming language into the target source program described in the second programming language, the clue information of the target source program is inserted. A program conversion program for causing a computer to execute a source program generation procedure embedded in a location corresponding to a syntax tree based on the grammar of the second programming language .   The program conversion program according to claim 1, wherein the source program generation procedure embeds the clue information in the target source program as a program comment. The source program generation procedure embeds a name of the starting source program and position information of the hierarchized data type in the starting source program as the clue information. Program conversion program.   The post-conversion model generation procedure is for data types that are included in the syntax tree of the starting source program and that do not have data types corresponding to the data types used in the first programming language in the second programming language. Is converted into a specialized data type created in advance so as to have a specification equivalent to the data type by the second programming language,
  In the target source program, when the specified data type is not used for comparison, the computer further executes a type conversion procedure for replacing the specified data type with a data type existing in the second programming language. The program conversion program according to claim 1, 2, or 3. A program conversion device for converting an origin source program described in a first programming language into a target source program described in a second programming language,
Pre-conversion model generation means for parsing the starting source program to generate a syntax tree of the source program based on the grammar of the first programming language;
A first storage unit for storing a conversion rule for converting a syntax tree based on the grammar of the first programming language into a syntax tree based on the grammar of the second programming language;
A second storage unit for storing a condition that the hierarchized data type is not subject to conversion based on the conversion rule;
When the syntax tree of the source program is converted into the syntax tree based on the grammar of the second programming language based on the conversion rule stored in the first storage unit, the syntax is stored in the second storage unit. The clue information indicating that the syntax tree based on the grammar of the second programming language corresponding to the syntax tree of the starting source program including the portion corresponding to the condition is excluded from the conversion based on the conversion rule. A converted model generation means to be inserted;
When converting the data model based on the grammar of the second programming language into the target source program described in the second programming language, the clue information of the target source program is inserted with the clue information A program conversion apparatus comprising: a source program generation unit embedded in a location corresponding to a syntax tree based on the grammar of the second programming language . A program conversion method for converting a source program written in a first programming language into a target source program written in a second programming language,
A pre-conversion model generation step of parsing the starting source program to generate a syntax tree of the starting source program based on the grammar of the first programming language;
Stored in the first storage unit, based on a syntax tree based on the grammar of the first programming language conversion rule for converting the syntax tree based on the grammar of the second programming language, the syntax tree of the origin source program Is converted into a syntax tree based on the grammar of the second programming language, based on the condition that the layered data type stored in the second storage unit is excluded from the conversion based on the conversion rule, The clue information indicating that the syntax tree based on the grammar of the second programming language corresponding to the syntax tree of the starting source program including the portion corresponding to the condition is excluded from the conversion based on the conversion rule. A post-conversion model generation process to be inserted;
A source program generating step of embedding the clue information in an appropriate portion of the target source program when converting a syntax tree based on the grammar of the second programming language into the target source program described in the second programming language; A program conversion method comprising: and.

Images