US 4,553,205 AGrant
Flexible Macro Expansion Process
Issue Date:1985-11-12
•8 Claims
•4 Drawing Sheets
Abstract
A method for conversion of digital computer source code software to operate in a system foreign to that from which the software originated. A library is created in which each entry incorporates the name of a code statement, or macro, which may require conversion and expanded code statements which may be selected to replace the macro. Each expanded code statement may be accompanied by one or more parametric triplets, each of which corresponds to a parameter within the expanded code and expresses a condition under which the expanded code is bypassed depending on the presence or absence of the parameter in any source code statement containing the macro. The source code statements are examined serially and any statement containing no macro entered in the library is passed to an output file, while any macro is replaced by any expanded code from the dictionary for which no bypass condition is satisfied. If it is desired to incorporate macros in expanded code, then the completed output file may replace the source file and the process may be repeated until no further macros are found.
Metadata
Inventor
- Salvatore Porchia
Application Information
Application Number:US 4207300
Filing Date:1982-09-21
Priority Date:1982-09-21
Art Unit:237
Classifications
IPC:
G06F 900G06F 1300
Field of Search:
364300;200;900
Patent Drawings (4 sheets)
Description
Background of the Invention
This invention relates to digital computer programming systems and, more specifically, to a method for conversion of software to operate in a systems environment foreign to that from which it originated.
The rapid growth of the computer hardware industry in the United States has placed on the market a wide variety of programmable digital computers at all price levels. The market for these machines has expanded rapidly because technological advances have caused the price of computing power to fall by a factor of 1,000 within the past decade. Thus, digital computers today are being applied to a far wider variety of tasks than before, including many tasks which, ten years ago, were more economical to perform by hand.
The proliferation of computers has resulted in a shortage of personnel trained to program these systems. This is reflected in the well-known high salaries currently offered to experienced programmers. Moreover, while organizations such as the American National Standards Institute (ANSI) have promoted programming language standards for use on all systems, most manufacturers produce a superset of the applicable ANSI standard language which amounts to a dialect applicable only to systems produced by that manufacturer. As a result, programmers who become familiar with one system and its dialect must spend substantial time learning a second dialect when writing software for a different system.
Also due to differences in dialect, software which has been written and debugged to run on a given machine will seldom, if ever, perform the same functions without error on a different machine. Timing constraints, file conventions, memory allocations and input-output control addresses, as well as other factors, all may be reflected in the text of a given software product. When attempting to modify software to operate on a system other than that from which it originated, failure to properly adjust these factors may cause not only loss of the software itself but also erasure of other files and peripheral storage.
As the shortage of programmers has become more widely recognized, and the libraries of useable software for particular systems have expanded, the need has arisen to convert software from one system to operate on a different system. This would allow programmer efforts to be directed to new developments rather than redevelopment of already-existing software. Generally speaking, however, where conversion has been attempted, it has required the effort of a programmer familiar with both the system from which the software originated and the system to which it is being converted. As noted above, such personnel are difficult to locate and command premium salaries. Therefore, there remains an uneconomical duplication and waste of scarce programming resources to produce software performing identical functions on different systems.
Summary of the Invention
The present invention provides a process for conversion of software to operate in a system of known dialect characteristics foreign to that from which the software originated. The invention examines the software source code for the presence of one or more statement portions which require change in order to function on the destination system. Such dialect statements are referred to as macros. As the source file is examined, an output file is created to which non-macro statements are transferred without change. Whenever a macro is recognized, the present invention allows it to be passed, deleted or replaced as appropriate to the destination system and dialect.
The present invention provides a flexible macro recognition system which may be easily modified to accommodate the transfer of software between any two systems having known dialect characteristics. Likewise, the disposition of each macro, including the replacement text, if any, is easily altered for the same purposes.
The present invention contemplates that, while many macros may be operative dialect statements which will execute without modification on the system of origin, other statements may be general macros intended to signify a complex operation performed by a series of software statements. It has been determined that the manipulation of data records may be advantageously divided into five basic functional activities, namely merge, update, extract, summarize and sort, all of which may readily be expressed as macros in the environment of a uniform discipline of file definitions and conventions. This permits a programmer to draft idealized or generalized softward using macros initially, followed later by preparation of the specific macro replacement statements necessary for each destination system dialect. Moreover, the statements which replace a macro may themselves advantageously include other macros. Therefore, the present invention contemplates that each completed output file may be treated as an input file and the process repeated, until an input file is found to have no remaining macro present therein.
A library file contains entries, each of which comprises the name of a macro, one or more expansion code statements which may be utilized to replace the macro in an output file, and one or more parametric triplets. Each expansion code statement may contain one or more parameters, with each parametric triplet corresponding to a parameter within the expansion code. The triplet signifies the disposition of the expansion code statement based on the presence or absence of the corresponding parameter in the original source code statement.
It can be readily appreciated that the present invention therefore provides a rapid and efficient method and apparatus for conversion of existing software products from one system to another. The invention may be used repeatly once the dialect characteristics of the system of origin and the system of destination are known and defined. The invention relieves the diversion of scarce and highly-paid programmer personnel from the task to which such personnel are best suited, namely the creation of new and original software products.
Although the invention itself is set forth in the appended claims, these and other objects and advantages of the present invention may be more readily appreciated by reference to the following detailed description thereof when read in conjunction with the accompanying drawings.
Brief Description of the Drawings
FIG. 1 is a flowchart of a portion of the inventive process;
FIG. 2 is a flowchart of a portion of the inventive process which occurs subsequent to that shown in FIG. 1;
FIG. 3 is a flowchart of a conditional portion of the inventive process, executed with certain conditions arise in the process of FIG. 2; and
FIG. 4 is a flowchart of another conditional portion of the inventive process which is executed when certain situations arise in the process in FIG. 2.
Detailed Description of the Drawings
The best mode and preferred of the present inventive process is illustrated in FIGS. 1 through 4. The inventive process, generally indicated at 10, utilizes logical directives expressed in a library file to transform a source file into an output file. Each of these files may be accessed sequentially, in any manner well known in the art.
The source file comprises a plurality of records, and each record contains one code statement. The file may be bounded by a beginning-of-file marker and an end-of-file marker. Each code statement may be in any recognized language. Among the text of each statement may be located one or more strings of characters comprising a system-dependent code, which will require alteration in order to operate as intended on a system foreign to the system of origin. Such a string of characters is referred to as a macro. Additionally, a macro may require replacement in some instances only on the presence or absence within the same statement of other strings of characters, referred to as parameters. Thus, an individual statement within the source file may contain no macro, or may contain a macro without parameters, or may contain a macro with parameters.
The library file may be subdivided into groups of records, each of which is referred to as an entry. The number of records per entry may be varied. Each entry is bounded by a beginning-of-entry indicator and an end-of-entry indicator. The records within each entry comprise strings of characters including a macro name, a library function indicator, and one or more model statements. Each model statement comprises two fields, one of which comprises an expansion code statement and the other of which may comprise one or more logical expressions here referred to as parametric triplets. The expansion code statement may incorporate one or more parameters, each of which correspond to a parametric triplet. As will be explained in greater detail below, each triplet is a logical expression utilized to determine whether the expansion code statement will be passed to the output file, based on the presence or absence of the corresponding parameter in the original source statement including the macro.
The output file comprises records containing statements transferred sequentially from the source file, intermixed with expansion statements from library entries in replacement of macros and, as required, parameters.
It should be noted that the inventive process and the library file may be expressed as read-only memory (ROM) structures in a general-purpose programmable digital computer.
The inventive progress begins at step 13 with a determination, typically by user input, of the function code to be utilized for library entry selection. As will be seen below, this allows the process to select among different entries, corresponding to different destination systems, for the same macro. Another use is to select entries that expand macros with documentation comments, thus automating the process of creating source file documentation. A further use is to select macros that will specify and define data files and field labels according to a standard structure and discipline to enable the use of file manipulation macros in the source code file. Yet another use is to select macros that will generate initial code for a system of known dialect rather than translate existing code to a destination system.
At step 15, initialization of the source file occurs which, as is well known, may comprise directing a source pointer to indicate the beginning-of-file marker. The pointer typically comprises a numeric value representing the position of the indicated record in main memory or, in a disk operating system environment, on disk. As each record is sequentially fetched, the pointer is incremented by an appropriate value corresponding to the length of each record.
Throughout this specification, each step of the inventive process as illustrated in the figures may be referred to by the numeral which labels the corresponding flow chart subdivision in the figures. The destination of process control transfer is illustrated by an encircled numeral label, which represents a branch to the process step bearing the label.
At step 18, the macro flag pointer is cleared. Following step 18, the next source statement is fetched at step 20, and is examined at step 30 to determine if the end of the source file has been encountered. If so, then the macro flag is examined to determine if it remains cleared. If this condition is also true, then no macros have been located in the source file and the job is at an end 34. If the macro flag has been set, the output file may optionally be passed to the source file at step 36, and step 15 and subsequent steps are conducted again until the source file no longer contains macros.
If the end of the source file is not encountered at step 30, then the library file is initialized at step 40 and the next library entry is fetched at step 50. At step 60, if the end of the library file has been reached, control is transferred to step 66 where the source statement is passed to the output file, and then to step 20. If the end of the library file has not been reached, the library entry fetched in step 50 is examined at step 62 to determine whether the function code designated in step 13 matches the function code of the library entry. If no match is found, control is transferred to step 50. If a match is found, the library entry is again examined at step 54 to determine if the macro name within the entry is found within the source statement fetched at step 20. If no match is found, control is again transferred to step 50.
Upon determination at step 64 that the macro name in the library entry appears in the current source statement, then the first model statement of the dictionary entry is fetched by initialization at step 70 and a fetch at step 80. Each model statement comprises two fields, one of which incorporates expansion code statements which may include one or more parameters, and the second of which may incorporate one or more parametric triplets. Each triplet identifies a corresponding parameter within the expansion code statement field. The triplet field may remain empty.
At step 90, if the end of model statements has been reached, control is transferred to step 20 at which the next source statement is then fetched. If the end of model statements is not reached at step 90, control is transferred to step 100 at which the triplets are initialized, followed by step 110 at which the next triplet is fetched. At step 120, if the triplets in the current model statement have been exhausted, control is transferred to step 260.
The expanded code field is then fetched and, at statement 280, transferred to the output file. At statement 285 the macro flag is then set, thereby indicating that at least one expanded code field has been inserted in the output file. Control is then transferred to step 80.
If the triplets have not been exhausted at statement 120, control is transferred to statement 130.
Each triplet is preferrably a text string comprising indication that the string is a triplet, indication of the parameter corresponding to the triplet and an indication of the replacement mode to be exercised. For instance, in an exemplary model statement, parameters may be stated as text string separated by delimiters, in the expanded code field. The first character of each triplet would indicate the status of the string as a triplet, the second character would specify the sequential position of the corresponding parameter within the code field and the third character would specify the replacement mode to be exercised. However, the present invention contemplates the use of any symbolic expression of these three attributes of a triplet, and is not limited to the use of a discrete character reflecting any one of the attributes.
Following statement 120, the first replacement mode to be tested may be unconditional replacement at step 130. For this mode alone, the triplet may appear without a corresponding parameter in the model statement. If unconditional replacement is indicated by the triplet, control passes to step 260.
If the triplet does not indicate unconditional replacement at step 130, at step 140 the triplet is tested for indication of AB mode. If this mode is indicated, then control passes to step 240 at which an incremental symbolic address is generated. At step 250, this symbolic address is passed to the output file, and control is then passed to step 260. Thus, the expanded code statement in the output file will be labelled with the symbolic address generated at step 240.
If AB mode is not indicated at step 140, control is transferred to step 150 where the triplet is tested for indication of A mode. If this mode is indicated, control is transferred to step 170 where the model statement parameters are initialized, and statement 180 where the parameter corresponding to the triplet is fetched. At step 190, the original source statement is then examined for the presence of the corresponding parameter. If the parameter is found in the original source statement, then control is transferred to step 80 which bypasses transfer of the expansion code from which the parameter was drawn in step 190. If the parameter was not present in the source statement at step 190, then control is transferred to step 110 and the next triplet is fetched for processing.
If the triplet is found not to indicate A mode at step 150, control is transferred to step 160 where the triplet is tested for indication of B mode. If this mode is indicated, control is transferred to step 200 where the expansion code field parameters are initialized, followed by step 210 where the triplet parameter is fetched. At step 220, the original source statement is examined for the presence of the fetched parameter. If the parameter is not present, control is transferred to statement 80. If the parameter is present, then control is transferred to statement 110. Thus, it is clear that the process steps conducted in B mode are substantially the inverse of those conducted in A mode.
If the parametric triplet does not indicate B mode at statement 160, it is an undefined triplet, and control is transferred to statement 110 at which, as noted above, the next triplet in the triplet field is fetched.
The present inventive process has been implemented on a NCR-IMOS computer in the COBOL language. A listing of this implementation follows, in which certain preferred forms are utilized for purposes of illustration only and not by way of limitation of the present invention. The process itself is expressed in COBOL, and the source file and library file are expressed in COBOL as well. Each parametric triplet within a model statement in a library entry is expressed in three consecutive characters. The first character is always the letter "Y", signifying the beginning of the triplet. The second character is numeric, to indicate the relative position of the parameter corresponding to the triplet within the expanded code field of the model statement. The third character of the triplet indicates the mode of the triplet. If the third character is numeric, then the mode is unconditional. If the third character is ASCII value 64 through 73, A mode is indicated. If the third character is ASCII value 74 through 83, then B mode is indicated. If the third character is ASCII value 84 through 93, then AB mode is indicated.
Certain other conventions are also observed in the working implementation. If the triplet field for a model statement is blank, a hyphen (-) placed in the 80th column of the model statement will cause recognition of this condition and placement of the expanded code field for the model statement in the output file. Otherwise, the triplet field is scanned from right to left and triplets are acted on as they are encountered. Each parameter within an expanded code field is delimited by a comma, and a maximum of ten characters are permitted per parameter, with a maximum of fifty parameters per library entry.
Four function codes are observed in the working implementation, all of which appear as a single-character prefix on the macro name in each library entry. The recognized characters are a slash (/), back slash (), period (.) and plus (+).
In the working implementation, each library entry includes a macro header record which includes the macro name and function code, the last up-dating date for the entry and comment fields. Following the model statements for the dictionary entry is an end record which consists solely of the word "END" and the macro name.
As noted above, the present inventive process may be advantageously utilized to implement and translate data file manipulation source code in terms of five basic functions: merge, update, extract, summarize and sort. Thus, each macro may be readily recognized in source code as having a specific basic function, and may be readily expressed in differing dialects within library entries. It can therefore be appreciated that the present invention enables a disciplined approach to the creation, documentation, maintenance and translation of data file manipulation source code.
Thus, the present invention provides flexible response to a wide variety of combinations of parameters appearing in an original source statement together with a macro. The decision chain represented by each parametric triplet field permits a precise definition of the conditions under which the associated expansion code field will be passed to the output field.
It should be noted that the steps of the inventive process can be interchangeable without departing from the scope of the invention. Furthermore, these steps can be interchanged and are equivalent. In the foregoing description, the invention has been described with reference to a particular preferred embodiment, although it is to be understood that the specific details shown are merely illustrative and the invention may be carried out in other ways without departing from the true spirit and scope of the following claims. ##SPC1## ##SPC2## ##SPC3##
Claims
What is claimed is:
1. A process for transformation of source code from an origin system of known characteristics for operation of a destination system of different known characteristics, with the aid of a digital computer, comprising the steps of: (a) providing said computer with a source file containing said source code, comprising a plurality of sequential source code statements, some of which incorporate at least one macro name and parameter, and with an output file; (b) providing said computer with a library file containing a plurality of sequential entries, each said entry comprising a macro name, a plurality of model statements, each said model statement comprising an expanded code field and a parametric triplet field, said expanded code field comprising expanded code which may incorporate parameters, and said triplet field comprising a number of triplets corresponding to the number of parameters in said expanded code field and each triplet identifying a parameter in said expanded code field and a condition of presence or absence of said parameter in said source code statements; (c) sequentially fetching a source code statement from said source file, and initializing said library file; (d) sequentially fetching an entry from said library file, and comparing said macro name of said entry with said source code statement; (e) repetition of step (d) until said library file is exhausted and, if no macro name is found in said source code statement, passing said source code statement to said output file and reversion to step (c); (f) sequentially fetching a model statement from said entry, if said macro name of said entry is found in said source code statement; (g) sequentially fetching a triplet in said parametric triplet field of said model statement, and a parameter identified by said triplet in said expanded code field of said model statement; (h) examining said source code statement for presence of said parameter; (i) repetition of steps (g) and (h) until said triplets are exhausted and, if each said condition identified by each said triplet is not satisfied, then passing said expanded code field of said model statement to said output file; (j) repetition of steps (f) through (i) until said model statements of said entry are exhausted; and (k) repetition of steps (c) through (j) until said source code statements are exhausted.
2. A process as claimed in claim 1 wherein each said sequential entry of said library file additionally comprises a function code provided in step (b), and further comprising the step after step (b) of providing said computer with a function code selection, and further comprising the step after step (d) of comparing each said function code of each said fetched entry with said function code selection and, if said function code of said entry is not equivalent to said function code selection, reversion to step (d).
3. A process as claimed in claim 1 including the step after step (b) of initializing and clearing a macro flag, the step after step (i) of setting said macro flag if said expanded code is passed to said output file, and the step after step (k) of passing said output file to said source file and repeating steps (a) through (k) if said macro flag is set.
4. A process of modification of source code developed for an origin system of known dialect, to produce source code for use in a destination system of differing from dialect, comprising the steps of: (a) providing a source file containing said source code, comprising a plurality of sequential source code statements incorporating parameters and macro names, each said macro name representing code to accomplish at least one basic function for data file manipulation including merge, update, extract, summarize and sort; and providing an empty output file; (b) providing said computer with a library file containing a plurality of sequential entries, each said entry comprising a macro name, one or more model statements, each said model statement comprising an expanded code statement which may incorporate one or more parameters, and a number of triplets each corresponding to one of the parameters in said expanded code statement, each said triplet specifying a condition of presence or absence of said corresponding parameter in a source code statement which inhibits passage of said expanded code to said output file; (c) sequentially fetching a source code statement from said source file, and initializing said library file; (d) sequentially fetching an entry from said library file, and comparing said macro name of said entry with said source code statement; (e) repetition of step (d) until said library file is exhausted and, if no macro name is found in said source code statement, passing said source code statement to said output file and reversion to step (c); (f) if said macro name of said entry is found in said source code statement, then sequentially fetching a model statement from said entry; (g) sequentially fetching a triplet in said model statement and a corresponding parameter in said expanded code statement of said model statement; (h) examining said source code statement for presence of said parameter; (i) repetition of steps (g) and (h) until said triplets of said model statement are exhausted and, if each said triplet does not inhibit passage of said expanded code statement to said output file, then passing said expanded code statement to said output file; (j) repetition of steps (f) through (i) until said model statements of said entry are exhausted; and (k) repetition of steps (c) through (j) until said source code statements are exhausted.
5. A process as claimed in claim 4 wherein said modes include a mode in which passage of said expanded code to said output file is inhibited by absence of said parameter in said source code statement.
6. A process as claimed in claim 4 wherein said modes include a mode in which passage of said expanded code to said output file is inhibited by presence of said parameter in said source code statement.
7. A process as claimed in claim 4 wherein said models include a mode in which passage of said expanded code statement to said output file is unconditionally conducted.
8. A method of transformation of source code, with the aid of a digital computer, comprising the steps of: (a) providing said computer with a source file containing said source code, comprising a plurality of sequential source code statements; (b) providing said computer with a library file comprising a plurality of sequential entries, each said entry comprising a macro name, a plurality of model statements, each said model statement comprising an expanded code field and a logical expression field, said expanded code field comprising expanded code which may incorporate a number of parameters, and said logical expression field comprising a number of logical expressions not greater than the number of parameters in said expanded code, each said logical expression indicating correspondence to one of said parameters in said expanded code statement and expressing a condition of appearance of said corresponding parameter in a source code statement also containing said macro name of said entry in which said logical expression is found, under which said expanded code replaces said source code statement; (c) examining each said source code statement for presence therein of any of said macro names; (d) if one said macro name is found in said source code statement, then examining each said model statement of said sequential entry of said library file containing said one macro name for presence of any said logical expressions and corresponding parameters, and examining said source code statement for presence of said corresponding parameters; and (e) if said source code statement fails to satisfy each said condition expressed by each said logical expression within one said model statement, then replacing said source code statement with said expanded code statement of said one said model statement.