JPS62221037A - Compiling method - Google Patents

Abstract

PURPOSE:To improve the efficiency of object production by applying the synthetic information on an intermediate text depending on a machine to an intermediate text obtained from a source program independent of the machine, and transforming the intermediate text into another depending on the machine to obtain an object program from the intermediate text. CONSTITUTION:An intermediate text is produced from analysis of the sentence structure of a source program independent of a machine. The synthesization enable attribute contained in the intermediate text is checked for display of another specific intermediate text with which the original intermediate text can be synthesized or not between the intermediate text. Then the intermediate text synthetic information depended on a machine is applied to the synthesization enable intermediate text for synthesization between the synthesization enable intermediate texts in response to a synthesization enable display. While on object program dependent on the machine is obtained from said synthetic intermediate text and those intermediate texts which are incapable of synthesization. Thus it is possible to output the compiled object program as a program dependent on the machine. Then the efficiency of object production is improved.

Description

[Detailed Description of the Invention] [Summary] Introducing a technique to perform as much synthesis processing as possible using machine characteristic information on intermediate text generated when compiling a machine-independent source program, thereby improving object efficiency. We aim to improve

[Industrial application field]

The present invention relates to a compilation method, and more particularly to a compilation method that improves the compilation of a machine-independent source program into a machine-dependent object program so as to improve the object efficiency.

Most recent program storage type data processing devices are configured to operate using programs written in a programming language, but in order to cause the device to actually operate, a source program written in a programming language is required. There is a need to convert it into an object program using a compiler.

Since the compiler is created depending on the target machine, problems arise when a source program created in a programming language is used on a different target machine.

[Conventional technology]

In other words, conventional compilers perform compilation processing that depends on the architecture of the target machine on which the object program generated by the compiler is executed. This dependence on architecture is not only reflected in the fact that the output object program differs depending on the target machine, but also in the analysis of the source program.

Therefore, for compilers with different target machines,
Generally speaking, the language specifications of the source programs that are applied to these programs are made to differ accordingly.

When the compilation method is adapted to such a source program, a problem arises in the distribution of the source program. In other words, the source program of one target machine can be reused on another target machine.

It is easy to make mistakes when you want to migrate or port, which is a big problem.

[Problem that the invention seeks to solve]

In order to avoid the problems that arise from changing the language specifications as described above, it would be sufficient to unify the language specifications, but such simple unification causes the following problems. That is,
Achieving the above-mentioned unification means that all compilers use a common syntax analysis unit; in other words, it means that analysis does not depend on the target machine. If an object program (assembler) is directly generated using this kind of analysis, the object program cannot take advantage of the characteristics of the target machine, which creates the dilemma of not being able to improve object efficiency. fall into

The present invention was created in view of such problems, and an object of the present invention is to provide a compiling method that can improve object creation efficiency and contribute to promoting distribution of source programs.

[Means for solving problems]

FIG. 1 shows a diagram explaining the principle of the present invention. As shown in this figure, the present invention includes a machine-independent source program syntax analysis step 10, and an attribute determination step 12 in which attributes that can be synthesized between the intermediate texts created in step 10 are determined and whether or not the synthesis is possible. and a synthesis step 14 in which machine-dependent intermediate text synthesis information is applied to the synthesizable intermediate text to synthesize synthesizeable intermediate texts according to the determination that synthesis is possible; and a code generation step 16 for generating an object program.

[For production]

Intermediate text is generated by machine-independent syntax analysis of the source program (step 10). Check the compositable attribute between the intermediate texts included in the intermediate text to determine whether the intermediate text can be composed with any other intermediate text, or with any other intermediate text of the intermediate text. An indication is made as to whether any combination of the two cannot be performed (step 12). Synthesis between the composable intermediate texts is performed by applying machine-dependent intermediate text synthesis information to the composable intermediate text in response to the composable intermediate text (
Step 14). A machine-dependent object program is generated from the synthesized intermediate text and the intermediate text determined to be impossible to synthesize in the above determination.

By doing this, even if the syntax analysis is common to all compilations, which can make machine-independent source programs, the resulting object program can be output as machine-dependent. This will improve object efficiency.

[Embodiment] FIG. 2 shows a processing flow of a main part of an embodiment of the present invention. Compilation based on this processing flow is performed in a system equipped with various addressing functions. Among the addressing functions, examples of addressing functions used in the compilation are as follows in relation to the compilation example described below.

1.1 Absolute address Description method: xxx XXX represents an absolute address.

Meaning: Represents data pointed to by an absolute address.

1.2 Address register indirect description method with displacement: d(An) d represents displacement (write a number).

An represents an address register.

Meaning: Represents data in memory that is pointed to by the sum of the contents of the An register and the displacement value.

1.3 Indexed address register description method: d (An, Ri) d represents displacement.

write numbers.

An represents an address register.

Rt represents the index register vinegar.

Meaning: Represents data on memory pointed to by the sum of the contents of the An register, the contents of the Ri register, and the displacement value.

The source program (in C language) compiled by the compiler of the present invention (see FIG. 3) before the processing according to the processing flow in FIG. 2 is started is as shown below, for example.

As shown in FIG. 3, this source program is analyzed by a machine-independent syntax analysis unit 20, and the intermediate text resulting from the analysis is stored in a file 30 in FIG. The intermediate text is as follows.

Flag ■ Find the start address of array ii 1 ■ Find the value 4 times n Add the values found in 1 ■■ and ■ 1 Add 4 to the value of ■■
Assign the content pointed to by the address of l■■ to variable i.
1'' indicates that the target machine supports the above-mentioned addressing function and can be reintegrated.

Next, processing according to the processing flow in FIG. 2 is performed.

This process in Figure 2 is performed by the intermediate text rewriting unit 4 in Figure 3.
0, this intermediate text rewriting unit 4
0, a determination is made as to whether the intermediate text read from the file 30 is the text for which an address is requested, and if the determination is negative, processing flow step 100, 1 in FIG.
After repeating the loop of 02 and 104, when it is determined that the next read intermediate text is the text for which an address is requested (determination is that it has the compositable attribute) (
(yes in step 104), the process proceeds to step 106. In the above stage classification, when the intermediate text is ■, the process proceeds to step 106.

In this step 106, the next intermediate text is read and it is determined whether the text is a text to which address information is to be added. Since the above-mentioned staged intermediate text ■, The double value (added value) and constant value 4 are temporarily held using a work area or the like.

Obtaining such a series of address information, that is, step 10
6,108,1) When the intermediate text read from the file is no longer the text to which address information is added while repeating the loop consisting of 0 (
Step 108 (no), it is determined whether the read intermediate text is a text using an address or not in Step Nob 1) 2, and if the determination is negative, step 10
6 (at this time, it is necessary that the flag is set to 1 in step 1). Flag - If ○, step 100
Proceed to. ), if the determination is affirmative, proceed to step 1)4.

The above setting intermediate text ■ corresponds to step 1) 2 that is yes, and in step 1) 4 intermediate text ■, ■
, ■ to form one intermediate text is performed as follows.

That is, the above-mentioned combination of the addressing functions that the above-mentioned compilation processing system is equipped with as described above, particularly the addressing function of the indexed address register indirect format determined by the above-mentioned address information and its intermediate text (machine-dependent intermediate text synthesis information) The process is activated to newly generate intermediate text ■°, and this is written in the file 30 as follows to rewrite it.

That is, the intermediate texts (2) and (2) are left as they are in the file 30, but the intermediate texts (2) and (2) are deleted from the file, and the intermediate text (2) is replaced with the intermediate text (3). In this way, the rewritten intermediate text is: ■ Find the start address of array ii ■ Find the value 4 times n When the intermediate text is substituted, the code creation unit 50 in FIG.
The assembler generated through
) Assign to register] ■' move 4 (ao, dO),
It becomes i.

Incidentally, the assembler generated by the code generation section 50 without performing the rewriting according to the present invention as described above is as follows.

As is clear from the above explanation, according to the present invention, machine-dependent intermediate text can be generated by the above-mentioned intermediate text rewriting process, so that the object program (assembler) generated therefrom takes into account the characteristics of the target machine. This can be used to improve object efficiency.

Other examples of the addressing function utilized in the embodiments of the present invention that have the above-mentioned effects will be added for the purpose of assisting in the disclosure of the present invention.

In the case of a source program to be run by the compiler of the present invention, the intermediate text obtained by being analyzed by the machine-independent syntax analysis unit 20 shown in FIG. Add the content that the value of ■■ points to. Add the constant 4 to the value of ■■. Assign the content that the value of ■■ points to to cc.
The intermediate text rewritten by the intermediate text rewriting unit 40 is as follows: ■ Find the content pointed to by the address a (displacement = 8). ■ Substitute the content pointed to by the address (displacement = 4) to cc. . This intermediate text■゛is an intermediate text■.

The addressing function of the absolute address format described above is used to obtain the intermediate text ■° from the intermediate texts ■ and ■, and the addressing function of the indirect format of the address register with displacement described above is used to obtain the intermediate text ■° from the intermediate texts ■ and ■.

The assembler that generated the intermediate text obtained in this way through the code generation section becomes , whereas the assembler that generates the intermediate text before rewriting according to the present invention through the code generation section 50 becomes [Effects of the Invention] ] As explained above, according to the present invention, machine-dependent intermediate text synthesis information is applied to intermediate text obtained from a machine-independent source program to transform the intermediate text into machine-dependent intermediate text, and the intermediate text is transformed into machine-dependent intermediate text. Since the object program is obtained from the text, the object creation efficiency can be improved.

[Brief explanation of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention, FIG. 2 is a processing flow of a main part of an embodiment of the present invention, and FIG. 3 is a diagram showing a compiler configuration of the present invention. In FIG. 1, 10 is a syntax analysis step, 12 is an attribute determination step, 14 is a synthesis step, and 16 is a code generation step. Source program Honmei principle explanatory diagram Figure 1 Honkomei compiler configuration diagram Figure 3

Claims (2)

[Claims] (1) Parse the machine-independent source program (1)
0), determines the compositable attribute between the intermediate texts of the syntax analysis results, displays whether or not the compositable is possible (12), and applies machine-dependent intermediate text synthesis information to the compositable intermediate text according to the determination of compositable. (14), and generates a machine-dependent object program from the synthesized intermediate text and the unsynthesized intermediate text (14).
16) A compilation method characterized by the following. (2) The address that constitutes the intermediate text synthesis information and is output for the synthesizable intermediate text is used by the addressing function of the system that supports the compile function when composing the synthesizable intermediate text. A compilation method according to claim 1, characterized in that: