JPH01166133A - Compile processing system - Google Patents

Abstract

PURPOSE:To shorten the access of a dictionary in a back end processing and to improve translation performance by setting the dictionary prepared in a front end processing in an internal common work area in the back end processing. CONSTITUTION:First, in a front end processing, a source program SP is read and the source program SP is analyzed. As the result of the analysis, an intermediate text 20 and a dictionary 21 are prepared in an external work area 2a. Next, in a back end processing 3, the intermediate text 20 is read by an input module 3a, the dictionary 21 in the external work area 2a is read and set in an internal common work area 3b. By the other processing modules 30 and 31 to prepare an object program OP, the object program OP is successively prepared while referring to the dictionary in the internal common work area 3b to the read intermediate text 20.

Description

[Detailed description of the invention] [Table of contents] Overview Industrial field of application Prior art (Figures 4 and 5) Problems to be solved by the invention Means for solving the problems (Figure 1) Functional example (a) Description of one example (Figures 2 and 3) (b)
Description of Other Embodiments Effects of the Invention [Summary] This invention relates to a compile processing method in which translation processing of a source program is performed by front-end processing and back-end processing.

The purpose is to reduce dictionary access time in back-end processing and improve translation performance.

A compilation processing method that performs front-end processing that analyzes a source program and creates intermediate text and a dictionary, and then performs back-end processing that generates an object program using the intermediate text and dictionary stored in an external work area. , in the back-end processing, the input module that reads the intermediate text reads the dictionary and sets it in the internal common work area, and other processing modules for generating the object program refer to the internal common work area and set the dictionary. Get the content.

[Industrial application field]

The present invention relates to a compile processing method that performs translation processing of a source program using front-end processing and back-end processing.

Converting a source program (original program) into a machine language object program is necessary in order for a computer to execute a program written in a simple language program, and for this purpose, compilers are used. The compiler is divided into preprocessing (front-end processing) and post-processing (back-end processing) so that it can generate object programs suitable for the target machine.

A compiler with such a configuration can generate object programs for nine different target machines, such as general-purpose computers, office processors, and workstations, simply by changing the back-end processing to match the target machine.

It's convenient.

It is desired to reduce the translation time in the back-end processing of such a compiler.

[Conventional technology]

FIG. 4 is an explanatory diagram of the prior art.

In the figure, SP is the source program, ifi, C0B
It is a primitive program written in a language such as OL,
OP is an object program (module)9
It has been translated into machine language.

1 is done by a compiler; in reality, the data processing system performs a translation program. 2) Lexical analysis is performed by a data processing system.

It is divided into front-end processing 2 that performs syntax analysis and semantic analysis, and back-end processing 3 that generates objects.

In the front-end processing 2, the source program SP is read, subjected to bilexical analysis, syntactic analysis, and semantic analysis, and the source program SP is converted into intermediate text 20 having a more concise structure and output.

For example, as shown in Figure 5, "Multiplying B and C is the combined product and D
The source program description that means "storing the sum of l'-B and C in work (WK)" means "storing the sum of work and D in A". is converted into an intermediate text element called "Storing".

In addition, in the front-end processing, a dictionary 21 is created for each data written in the source program, and the data (A, B in FIG. 5) is stored in the dictionary.

For attributes (number of digits, etc.) in C and D), information such as addresses is set.

On the other hand, in the "hack en)" process j13, the intermediate text 20 generated by the front end is read, and objects are successively generated by the processing modules 30 to 3n.

During this process, when you want to know the attributes of the data.

Refer to V21.

Such intermediate text20. The amount of dictionary 21 created is variable depending on the number of steps in the source program and how it is written.

Therefore, an area cannot be fixedly allocated to the compiler internal work area, but is created in the external work area 2a.

[Problem that the invention seeks to solve]

However, in the conventional technology, since the dictionary 21 is created in the external work area 2a of the compiler, the back-end processing 3
In order for each of the processing modules 30 to 3n to refer to the dictionary 21, it is necessary to access the external work area one by one by processing such as file opening, which poses a problem in that it takes a long time to access.

Therefore, a considerable amount of time is required just for access processing that is not directly related to translation, reducing the translation performance of the compiler.

An object of the present invention is to provide a compilation processing method that can shorten dictionary access time in back-end processing and improve translation performance.

[Means for solving problems]

FIG. 1 is a diagram explaining the principle of the present invention.

In the figure, the same parts as those shown in FIG. 4 are indicated by the same symbols.

In the present invention, in the back-end processing 3, the input module (3a) that reads the intermediate text 20 reads the dictionary 21 in the external work area 2a, sets it in the internal common work area 3b, and generates an object program (OP). Other processing modules 30 and 31 are in the internal common work area 3b.
The contents of the dictionary are obtained by referring to the .

[For production]

In the present invention, the input module 3a reads a dictionary corresponding to the operand appearing in the intermediate text.

Necessary information is set in the common work area 3b inside the compiler.

This internal common work area 3b can be referenced in common from any module that makes up the compiler, so the processing modules 30.31 do not have to access the external work area 2a.
You can refer to necessary dictionary contents such as data attributes. - Therefore, access time can be significantly reduced and translation performance can be improved.

〔Example〕

(a) Description of an Embodiment FIG. 2d is a detailed explanatory diagram of the present invention. FIG. 3 is a diagram showing the relationship between the intermediate text and the dictionary in FIG.

In Figure 2, the same parts as those shown in Figures 1 and 4 are indicated by the same symbols. 4 is a source program file fi, source written in C0BOL language etc. 5 is an object file in which the program is stored; 0 is the object file that stores the generated object program; 0 is the processing module 30, which is an arithmetic processing module; , the processing module 31 is a transcription processing module that creates an object for transcription processing.

Hereinafter, as a source program SP, B and C are added and stored in A [cOMPUTE A=B+CJ as the code part and data A as the data part.

An example of the C0BOL language that describes the number of digits (PIC) of B and C will be explained.

■ Source program SP of source program file 4
as input to front end path processing 2.

Start front-end path processing.

■ As the output result of the compiler's front-end pass processing 2, an intermediate text file 20 and a dictionary file 21 are created.
is generated.

As shown in FIG. 3, an instruction identification code (ADD) and data A and B are stored in the intermediate text file 20.

A dictionary reference value of C is generated, and the dictionary reference value means the relative position from the beginning of the dictionary file 21 to the target element.

On the other hand, the dictionary file 21 contains data A and B.

0 where the elements of C (number of digits, number of digits in the second place, etc.) are stored ■ The generated intermediate text file 20 and dictionary file 21 are input to the back-end pass processing 3 of the compiler, and the back-end pass is started. Yes 0 ■ As the start process of backend path process 3.

The intermediate text human module 3a reads the intermediate text from the intermediate text file 20.

Then, from the dictionary reference values of names A, B, and C,
Elements related to each of A, B, and C are read from the dictionary file 21.

Then, values are set in the operand prototypes of A, B, and C, and the completed operands are set in the compiler internal common work area 3b.

The operand prototype is a format that can be commonly used by the following modules. '■ Next, the arithmetic processing module 30 is called.

The arithmetic processing module 30 refers to necessary information such as the operand prototype created by the intermediate text input module 3a and the attribute information, and generates an object that adds the value of -will B and the value of will C) rMOVAX, - will BJ, l'
-ADD AX, -Output will cj to object file 5.

After that, the AX register in which the calculation result (B+C) is stored is transferred to the unique data area.

The transcription processing module 31 is called.

(2) The transcription processing module 31 outputs to the object file 5 the object [MOv-Will A, AXJ for transcribing the AX register to the unique data area.

At this time, the attribute information of the person required to create the object is obtained by referring to the operand prototype of the unique master created by the intermediate text human power module 3a.

In this way, the intermediate text manual module 3a only needs to access the dictionary file 21 once.

The other lower module modules 30 and 31 can determine necessary information such as the attribute information of -will without accessing the dictionary file 21.

Therefore, by reducing the number of times the dictionary file 21, which is the external work area 2a, is accessed, the time spent accessing the dictionary file 21 is saved, and translation performance is improved.

Further, since the intermediate text input module 3a sequentially reads the intermediate text, the dictionary contents necessary at that time are read into the internal common work area 3b, so the two work areas 3b may be fixed, so they are provided internally.

(b) Description of other embodiments In the embodiments described above, the C0BOL language was used for the source program, but other languages may be used.

The source description is not limited to the one shown in FIG. 2, but can also be applied to other sources.

Also, the processing module was explained in terms of calculation and transcription, but
Various modules necessary for generating objects can be used.

Although the present invention has been described above with reference to embodiments, the present invention can be modified in various ways according to the spirit of the present invention.

These are not excluded from the present invention.

〔Effect of the invention〕

As explained above, according to one aspect of the present invention, the number of accesses to the external work area during back-end processing can be reduced in a system in which compilation processing is performed by front-end and back-end processing.

This has the effect of improving translation performance and speeding up compilation speed.

[Brief explanation of the drawing]

FIG. 1 is a diagram explaining the principle of the present invention. FIG. 2 is a detailed explanatory diagram of the present invention. FIG. 3 is a diagram showing the relationship between the intermediate text and the dictionary in FIG. 2. FIG. 4 is an explanatory diagram of the prior art. FIG. 5 is an explanatory diagram of front end processing. In the figure, 1...compiler. 2...Front end processing. 2a...External work area. 3...Backend processing. 3a...input module. 3b... Internal common work area. SP...Source program. OP...Object program. 21...Dictionary file. Patented rice cake Fujitsu Limited

Claims (1)

[Claims] A source program (SP) is analyzed, intermediate text (2
Front-end processing (2) that creates a dictionary (21) and a dictionary (21)
), the intermediate text (20) stored in the external work area (2a)
and dictionary (21) to create an object program (O
In a compilation processing method that performs back-end processing (3) to generate P), in the back-end processing (3), an input module (3a) that reads the intermediate text reads the dictionary and creates an internal common work area (3b). , and another processing module (30) for generating an object program (OP) refers to the internal common work area (3b) to obtain dictionary contents.