JPH0628165A - Automatic code diagnostic device based upon coding rule - Google Patents

Abstract

PURPOSE:To provide the general device which automatically diagnoses code to be inputted to a computer after coding according to a specific grammar based on the separately defined coding rule. CONSTITUTION:This device is provided with a means 108 which decides specific conditions to be diagnosed by the specification of rule information, a means 110 which processes information to be diagnosed by the specification of the rule information, a means 111 which checks conditions that the object of diagnosis should satisfy by the specification of the rule information, and a rule control means 107 which calls the three means 108, 110, and 111 independently of one another according to the rule information. Further, this device includes a pattern check means which checks whether or not the codes satisfy rules regarding coding patterns by checking whether the codes and coding patterns match each other while comparing the rules regarding them with each other, character by character.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention makes a command (hereinafter referred to as a code) input to a computer such as a program source or a JCL (Job Control Language) source coded according to a certain grammar into a separately defined coding rule. The present invention relates to a device that automatically diagnoses based on the above.

[0002]

2. Description of the Related Art In addition to basic grammatical checks by a compiler or assembler, a program is designed to be more logical and effective before it is actually executed on a real machine for the purpose of efficient debugging and effective use of resources. It is effective to check from a high point of view such as maintainability. In order to improve the maintainability of programs, it is necessary to standardize coding, and it is important to check the coding style of created programs.

Performing such a check manually by a human is easy to overlook and takes a lot of time. Therefore, some automatic diagnosis devices have been proposed. For example, JP-A-63-12027
JP-A-63-263545, JP-A-63
No. 280344, JP-A-3-5837 and JP-A-3-288226. These devices are externally stored rules (coding rules expressed in a form that the diagnostic device can interpret).
Based on the above, the description format of the code of the program etc. is automatically diagnosed, and the rules are externally stored so that the rules can be freely rewritten. Further, it has been proposed to define the rule expression by using a simple language so that it can be easily changed.

[0004]

However, although the above-mentioned prior art device has means for registering rules externally in order to increase versatility and inputting them at the time of check processing, It is not a deep study of how to express that rule. In other words, since there was no mechanism for checking complicated coding rules, it was not possible for the user to check all the rules that he would like to check without recreating the code automatic diagnosis device, but within a limited range. Only minor changes and extensions to the rules have been made. This problem is practically important, and if the function provided for expressing the rules is not sufficient, the necessary rules cannot be expressed, and as a result, the know-how of the experienced programmer cannot be completely described. As described above, the code automatic diagnosis device having a low rule changing / expanding ability cannot express the coding rules necessary in the actual work, and thus has not been fully utilized.

Further, in the prior art, as a check of the coding pattern (character face of the code described), there is a function for indentation / indentation return and a check in the case where a plurality of instructions are described in one line. It was not possible to strictly check the coding pattern without changing or expanding the diagnostic device. In the conventional code automatic diagnosis device, it is not possible to perform checks based on rules that strictly define the coding pattern such as the number of columns for indentation, the format of argument definitions in functions, and how to write comments up to the column position and the number of characters at which they are written. , Could not be implemented without expanding the device.

In order to solve the above problems, it is conceivable to incorporate means for checking a pattern in the code automatic diagnosis device. As a conventional technique for checking patterns, there is a pattern matching method for character strings using regular expressions. However, the regular expression only provided special characters such as "^" and "$" that represent the beginning and end of a line as a method of specifying the position of a specific character in a pattern. . These special characters only represent the beginning and end of a line, and cannot represent any column position in the line. Also, it is not possible to specify which line the line hits. In addition, "^" and "$"
When using, "^" is the beginning of the regular expression, "$"
There was a restriction that could only be used at the end of a regular expression.
Therefore, for example, the pattern "the characters" abc "are arranged from the 25th column of the line" is a regular expression: / ^. ． ． ． ． ． ． ． ． ． ． ． ． ． ． ． ． ． ． ． ． ． ． ． It had to be written as abc /, which was very annoying. In addition, for example, a rule that specifies an absolute line position such as "this description should come in the first line" could not be expressed.

In addition, in the regular expression, the pattern cannot be expressed in consideration of the two-dimensional spread of the pattern of the character string, that is, the image as it is when the code is printed out or displayed on the screen. It was For example, if there is a coding pattern such as if △ ... △△ a △ ＝ ..., in the conventional regular expression, write /if△.\n△＊a△＝.*/ in one line. become. ("/" Represents the beginning and end of the regular expression, and "\ n" represents a line break. "△" or "
"_" Represents a blank space, and ". "" Represents an arbitrary character and "*" represents 0 or more repetitions of the preceding character.) With this, the image of the two-dimensional spread of the coding pattern disappears at all, and the regular expression is just hard to read. It cannot be expressed that the character "a" is located two columns lower than the "i" in "if." As described above, the conventional regular expression can be used to expand the two-dimensional spread of the character string pattern. There was a problem that it was not possible to express it, and if it was forcibly expressed in one line, the relative indentation could not be expressed.

Further, in the conventional regular expressions, the "paired description" cannot be checked. For example, in the coding pattern of {_a_ {x} _, _b {y} _} _ {_ c_} -------------, in the coding pattern of "{", "}"
Are used as a pair and can also be nested. Here, in order to match the underlined part of this coding pattern, a regular expression is used: / {_ a. When written as *} /, there was a case where it was matched with unpaired parentheses such as {_a_ {x} or {_a_ {x} _, _b {y} _} _ {_ c_}. As described above, in the conventional regular expressions, the description written to match the "paired description" often matches the "unpaired description". For this reason, the rule that was going to be checked could not be checked, and a situation in which misdiagnosis frequently occurred. And the human had to visually check the misdiagnosis. This was a very time-consuming task.

Further, in the conventional regular expression, it is not possible to reuse the separately defined pattern in the regular expression. Therefore, the same coding pattern had to be described in the rule many times, which was inefficient.

In view of such circumstances, the first object of the present invention is to
It is an object of the present invention to provide a general-purpose code automatic diagnosis device that allows a user to check a rule to be checked without recreating the code automatic diagnosis device or adding a means for checking the coding rule.

A second object of the present invention is to set coding patterns such as "indentation format", "argument definition format in function", and "comment format" to the column positions and the number of characters described therein. When a user wants to check a rule that strictly defines everything, a general-purpose code automatic diagnosis that can be checked without recreating the code automatic diagnosis device or adding a means for checking the coding rule To provide a device.

A third object of the present invention is to automatically code a pattern rule by simply processing the pattern rule by designating the absolute position of a specific character in the pattern in checking the coding pattern. To provide a diagnostic device.

A fourth object of the present invention is to provide an automatic code diagnosing device capable of describing a pattern rule which is easy to write and easy to read, while paying attention to the two-dimensional spread of a character string pattern in checking a coding pattern. To do.

The fifth object of the present invention is to automatically check the rule to be checked by checking the "paired description" in checking the coding pattern, and to automatically check the code automatically. To provide a diagnostic device.

Further, a sixth object of the present invention is to provide an automatic code diagnosing apparatus which can improve the efficiency of rule description by reusing a separately defined pattern in checking a coding pattern. is there.

[0016]

To achieve the first object, according to the present invention, a code coded according to a predetermined grammar and input to a computer is automatically generated based on a separately defined coding rule. In the apparatus for diagnosing automatically, the code automatic diagnosing apparatus is configured as follows. That is, the code automatic diagnosis device of the present invention includes a rule input means for inputting a rule expressing the coding rule in a form interpretable by the device and storing rule information in a rule DB, and analyzing the code to generate a code. Code analysis means for storing information in the code information DB, and a "diagnosis target" for determining whether or not a specific position of the code satisfies a diagnosis target specifying condition that is a condition for specifying the diagnosis target specified in the rule information. "Means for determining specific conditions", "Means for processing diagnosis target information" for executing "processing of diagnosis target information" specified by the rule information, and "Satisfaction target for diagnosis specified by the rule information""Means to check the conditions that must be satisfied by the diagnosis target"
Rule control means for independently calling the "means for determining a diagnosis target specific condition", "means for processing diagnosis target information", and "means for checking conditions to be satisfied by the diagnosis target" according to the rule information; A storage unit for writing and referring to information in a processing process is provided by the "means for determining a diagnosis target specific condition", "means for processing diagnosis target information" and "means for checking conditions to be satisfied by a diagnosis target", and And a "diagnosis message output means" which is called by the rule control means when the "condition to be satisfied by the diagnosis target" is not satisfied and which outputs a diagnosis message according to the rule information.

In order to achieve the second object, according to the present invention, there is provided an apparatus for automatically diagnosing a code coded according to a predetermined grammar and input to a computer based on a separately defined coding rule. , Including a "pattern checking means" for checking whether the code satisfies the rule for the coding pattern by checking whether the code matches the rule for the coding pattern character by character. A featured device is provided.

In order to achieve the third object, the code automatic diagnosis device of the present invention is configured such that the "pattern checking means" includes "absolute position checking means".
The "absolute position checking means" is a means for checking whether or not a character or a description included in a rule of a coding pattern is placed at a certain absolute position on a code.

In order to achieve the fourth object, the code automatic diagnosis apparatus of the present invention is configured such that the "pattern checking means" includes "means for controlling a coding pattern rule spread in two dimensions". . The "means for controlling a two-dimensionally spread coding pattern rule" is a method of checking a pattern rule while decomposing a two-dimensionally spread coding pattern rule line by line and correcting the relative position from the previous line. Is a means of performing.

In order to achieve the above fifth object, the code automatic diagnosis apparatus of the present invention is configured such that the "pattern checking means" includes "paired description checking means". The "paired description checking means" is called when there is a paired description in the pattern rule, and the paired description on the code and the paired description on the rule correspond to each other. Is a means of checking.

In order to achieve the sixth object, the code automatic diagnosis apparatus of the present invention is configured such that the "pattern checking means" includes "another pattern rule reference means". The "reference means of another pattern rule" is called when there is a "reference description of another pattern rule" in the pattern rule, fetches another pattern rule from the rule DB, and checks the pattern accordingly. It is a means for making it happen.

[0022]

In the automatic code diagnosing device configured as described above, "means for determining a diagnosis target specifying condition", "means for processing diagnosis target information" and "means for checking conditions to be satisfied by the diagnosis target". These three means are separated and each is independently called by the rule control means. For example, it is possible to repeat the "diagnosis target identification process" and the "diagnosis target information processing process" in multiple stages. You can Therefore, even complicated coding rules can be decomposed into a combination of "diagnosis target specific conditions", "processing of diagnosis target information", and "conditions that must be satisfied by the diagnosis target". , You can check it without adding any means to check the coding rules.

Further, since the "pattern checking means" is included, a means for recreating the code automatic diagnosis device or checking the coding rules for the coding pattern rules that strictly define the column position and the number of characters. You can check without adding.

Further, the "pattern checking means" includes "absolute position checking means", and it is possible to process the pattern rule designating the absolute position of a specific character in the pattern. Therefore, the pattern rule can be simply described. Further, it is possible to describe a coding rule that specifies an absolute line position of the description, which could not be described conventionally.

Further, the "pattern checking means" includes "means for controlling the coding pattern rule spread in two dimensions" and analyzes the pattern rule spread in two dimensions, so that the pattern rules are not described in one line. ,
It is possible to express the code as it is when it is printed out or displayed on the screen. Therefore, since the pattern rule is easy to write and easy to read, the description error of the coding rule is reduced and the correction is easy.

Further, the "pattern checking means" includes "paired description checking means" to prevent unmatched description from pattern matching. Therefore, an error caused by unmatched description pattern matching is prevented. The diagnosis can be eliminated. Therefore, a lot of time spent for the erroneous diagnosis check can be saved.

Further, the "pattern check means" includes "another pattern rule reference means", and the separately defined pattern can be reused in the coding pattern check. Therefore, the efficiency of rule description can be improved and correction can be easily performed.

[0028]

Embodiments of the present invention will now be described in detail with reference to the accompanying drawings. An automatic code diagnosis apparatus according to an embodiment of the present invention is a microcomputer / computer as shown in FIG.
It is embodied in a system and software running on it. In the following description, the storage means is implemented by the main storage device 602 shown in FIG. 17, and the diagnostic message output means is implemented by the printer output device 604. Further, the input function of the rule input means and the code analysis means is embodied by the input / output terminal device 603, and the analysis function is the CP function.
It is embodied by software running on U601.
Further, the rule DB and the code information DB are created in the external storage device 605. Each of the other means is realized by software that runs on the CPU 601 and that realizes its function.

First, description will be made with reference to FIG. Figure 1
It is a figure which shows the structure of the code automatic diagnosis apparatus of this invention. In FIG. 1, the rule 101 is the rule input means 10 in advance.
3 is input and stored in the rule DB 104. Further, the code 102 is analyzed by the code analysis means 105, and its information (code information) is code information DB10.
6 is stored. The code analysis means 105 here includes a syntax analysis means and a clause analysis means.

Next, the rule control means 107 causes the rule D
The rule information is read from B104. Then, the rule control means 107 calls the means 108 for determining the diagnosis target specific condition, the means 110 for processing the diagnosis target information, and the means 111 for checking the conditions to be satisfied by the diagnosis target according to the rule information. Here, these three means 108, 110, 111 fulfill their respective functions with reference to the three pieces of information of the code 102, the code information DB 106, and the rule DB 104. Further, the information is written in the storage means 109, and then the information is referred to.

The rule control means 107 uses the "diagnosis target specific condition" in the rule information and the "absolute position of the code" (information specifying the row and column position on the code) as "means for judging the diagnostic target specific condition". Hand it over to 108. The “means for determining the diagnosis target specifying condition” 108 determines whether the description of the code starting from the designated “absolute position of the code” satisfies the diagnosis target specifying condition. When the condition is satisfied, information such as position information obtained at the time of determination (“specified diagnosis target information”) is written in the storage unit 109.

The rule control means 107 passes the "processing for processing the diagnosis target information" in the rule information to the "means for processing the diagnosis target information" 110. The “diagnosis target information processing unit” 110 writes the information (diagnosis target processing information) generated as a result of executing and processing the “diagnosis target information processing process” to the storage unit 109.

The rule control means 107 passes the "condition to be satisfied by the diagnosis target" in the rule information to the "means for checking the condition to be satisfied by the diagnosis target" 111. The “means for checking conditions to be satisfied by the diagnosis target” 111 is based on the specified diagnosis target information, diagnosis target processing information, information of the code information DB 106, and information of the code 102.
Check the "conditions that the diagnostic target must meet".

If, as a result of this check, the "condition to be satisfied by the diagnosis target" is not satisfied, the rule control means 107 passes the diagnosis message in the rule information to the "diagnosis message output means" 112. The “diagnostic message output means” 112 receives the diagnostic message,
The information in the storage means 109 is edited and output on a form or screen (113).

FIG. 2 is a diagram showing the internal structure of the "means for judging the diagnosis target specifying condition" 108 of FIG. In FIG. 2, as for the external components, only those importantly related to them are shown. It does not mean that there is no connection just because it is omitted, and in particular, the storage means 109 is omitted because it is related to all the means in FIG.

[Means for Determining Diagnostic Target Specific Condition] 10
The "diagnosis target specific condition" received by the 8 is separated into the pattern rule and the attribute condition that the syntax element should satisfy by the first logic determining means 201, and the pattern checking means 501 and the attribute that the syntax element should satisfy, respectively. It is passed to the condition checking means 202. The pattern checking unit 501 determines whether the “absolute position of the code” satisfies the pattern rule. Hereinafter, this process is called a pattern check. A state in which the pattern rules are satisfied by the pattern check is called a state in which the patterns match. Since the pattern rule generally includes syntax information, it is necessary to refer to the code information DB 106 for checking. The attribute condition check means 202 to be satisfied by the syntax element is activated only when the pattern matches,
Check the attribute conditions that must be satisfied by the syntax elements included in the pattern rule. At this time, the code information DB 106 is referred to.

In this way, the pattern rule and the "attribute condition that the syntax element should meet" are checked by the "pattern checking means" 501 and the "attribute condition check means that the syntax element should meet" 202. The result is returned to the first logic judgment means 201. The first logic determining unit 201 determines whether the diagnostic target specific condition is true or false based on the pattern rule and the true / false value of the “attribute condition that the syntax element should satisfy”, and returns the result to the rule control unit 107.

FIG. 3 is a diagram showing the internal structure of the "means for processing the diagnosis object information" 110 of FIG. In FIG. 3, regarding the external components, only those importantly related to them are shown. The fact that it is omitted does not mean that there is no relation, and especially with the storage means 109,
It is omitted because it is related to all the means in FIG.

The "processing for processing the diagnosis target information" received by the "processing means for processing the diagnosis target information" 110 is separated into the range information extraction processing and the DB operation processing by the processing distribution means 301. Information extraction means 3
02, the DB operating means 303. The range information extracting unit 302 includes a range of partial patterns included in the pattern rule, an innermost nest range including the pattern rule,
Is a means for extracting range information. The nesting range is, for example, a range surrounded by “{” and “}” that form a pair in C language. It has a multi-layered structure (one hierarchy may be better) where one nesting scope contains another nesting scope. The range information for the syntax element is stored in the code information DB 106 and needs to be referred to. The DB operating means 303 is a means for operating the information (stored as a table) stored in the code information DB 106 and storing the information processed by the DB operation in the storage means 109 as a new table.

FIG. 4 is a diagram showing the internal structure of the "means for checking the condition to be satisfied by the diagnosis target" 111 of FIG. In FIG. 4, as for the external components, only those importantly related to them are shown. It does not mean that there is no connection just because it is omitted, and in particular, the storage means 109 is omitted because it is related to all the means in FIG.

The “condition to be satisfied by the diagnosis target” received by the “means for checking the condition to be satisfied by the diagnosis target” 111 is the pattern rule, the comparison condition, the arrangement condition of the table elements, And is passed to a pattern checking unit 501, a comparison condition checking unit 402, and a table element array condition checking unit 403, respectively. The pattern checking unit 501 indicates the position on the code specified by the diagnosis target specifying condition (hereinafter referred to as a pattern matching position. In the present embodiment, the “absolute position of the code” with which the first character of the pattern rule is matched). The pattern matching position) and the pattern rule are passed, and it is determined whether the diagnosis target satisfies the pattern rule.

The comparison condition checking means 402 is provided with the specified diagnosis target information, diagnosis target processing information, and code information DB.
This is a means for determining a comparison conditional expression regarding a numerical value or a character string consisting of information such as 106 information and code 102 information. Especially, in the comparison conditional expression, "NESTLEVE
When the special variable "L" appears, the "nesting level finding means" 401 is called to find the nesting level at the pattern matching position. The nesting level represents to which depth the nesting range a certain description on the code belongs to, and the nesting level of a part which does not belong to any nesting range is counted as 0. Table element arrangement condition check means 4
Reference numeral 03 is a means for checking whether certain elements (column data) in the table are arranged in ascending or descending order.

In this way, the pattern checking unit 501, the comparison condition checking unit 402, and the table element array condition checking unit 403 check the pattern rule, the comparison condition, and the table element array condition. No. 2 logic determination means 404. The second logic determination means 404 determines whether the “condition to be satisfied by the diagnosis target” is true or false based on the true / false values of the pattern rule, the comparison condition, the arrangement condition of the table elements, and the result is rule control means. Return to 107.

FIG. 5 is a diagram showing the internal structure of the pattern checking means 501, which is a component of FIGS. 2 and 4. In FIG. 5, regarding the external components, only those that are importantly related are shown. It does not mean that there is no connection just because it is omitted, and in particular, the storage means 109 is omitted because it is related to all the means in FIG.

In this embodiment, the "means for controlling the two-dimensionally spread coding pattern rule" is "means for converting a plurality of lines of coding pattern rules into one line".
6 and “relative position checking means” 505. "Means for converting a plurality of lines of coding pattern rules into one line" 506 decomposes a two-dimensionally spread coding pattern rule into lines, and inserts a special character that specifies a relative position from the previous line between the lines. It is a means to connect all of them together to make a single rule. This “means for converting a plurality of lines of coding pattern rules into one line” 506 can be included in the “pattern checking means” 501, and the rule input means 10
It can be included in 3. "Relative position checking means"
505 is whether the character or description following the special character that specifies the relative position included in the rule of the coding pattern is placed in the same column position as the column position where the pattern check on the code starts. Is a means of checking (row positions may be different).

First, "pattern checking means" 501
Receives two pieces of information, the absolute position of the code and the pattern rule. Information on the absolute position of the code is stored in the storage means 10.
9 is written. This position is called the pattern check start position. The pattern rule is passed to the “means for converting a plurality of lines of coding pattern rule into one line” 506. The “means for converting a plurality of lines of coding pattern rules into one line” 506 that received the pattern rules is
The pattern rule is converted into one line and passed to the “pattern rule control means” 502. "Pattern rule control means" 502 that received the pattern rule converted into one line
Means checking means 503 for matching one character, checking means 507 for paired description, checking means 50 for absolute position
4, relative position check means 505, syntax element description check means 509, different pattern rule reference means 50
The matching of the pattern rule and the code is checked by calling 8 respectively. The one-character matching check means 503 is a means for checking whether one character (a description representing) included in the pattern rule and one character on the code match. The syntax element description checking unit 509 is called when the pattern rule includes a syntax element description such as a "sentence" or a "conditional expression".
The description of the code to be checked is "statement" or "conditional expression"
Whether or not it corresponds to syntax elements such as code information DB
It is a means of checking by referring to.

When the special position checking means 504 finds a special character "\ nn" (nn is a two-digit integer) in the pattern rule, this "\ nn" of the pattern rule is detected.
Checks whether the position of the character in the code corresponding to the next character of is in the nnth column. When a special character "\! Nnnn" (nnnnn is a four-digit integer) is found in the pattern rule, it is checked whether the line including "\! Nnnn" in the pattern rule is the nnnn line. For example, "from the 25th column of the row" ab
The pattern rule saying "the characters" c "are lined up" is /. It can be simply written as * ¥ 25abc /. With the conventional regular expression, the line start and line end positions could be checked, but compared to this, there are the following advantages.・ You can specify not only the beginning and end of the line, but also the position within the line. -Absolute line position can be specified. -The absolute position can be specified not only at the beginning and end of the pattern rule but also anywhere in the pattern rule.

When the special character "\ #" is found in the pattern rule, the relative position checking means 505 determines that the column position of the character in the code corresponding to the character next to "\ #" in the pattern rule is Check whether it is the same as the column position of the pattern check start position.

The checking means 507 for the paired description is C
The language "{", "}" and "(", ")" are considered as paired descriptions, and when these descriptions appear in the pattern rule, they are called by the pattern rule control means 502. Upon receiving the paired description, the "paired description checking means" 507 first calls the one-character matching check means 503 to check whether the description in the code matches the paired description. If they match, the processing differs depending on whether the match is the description on the right side or the description on the left side. When the matched one is the description on the left side (in the case of "{", "("), the information in which the description on the left side in the code and the description on the left side in the pattern rule are associated with each other is written in the storage means 109. If the matched one is the description on the right side ("}", ")"), "the description on the left side that pairs with the description on the right side in the code" and "pairs the description on the right side in the pattern rule" It is checked whether the "description on the left side" is associated and written in the storage unit 109. As a result, it is possible to pattern match only the descriptions that form a pair.

However, "means for checking paired description"
Function of "correspondence of paired description on code", "means of correspondence of paired description on rule",
And, if it is decomposed into "a means for checking that the paired description on the code and the paired description on the rule correspond to each other", "a means for making the paired description on the code correspond" And “means for taking correspondence of paired description on rule” with respect to the code analysis means 10 respectively.
5, it may be included in the rule input means 103.

The different pattern rule reference unit 508 is called by the pattern rule control unit 502 when another rule is reused in the pattern rule. The different pattern rule reference unit 508 reads another rule used in the pattern rule from the rule DB 104, and reads “the pattern rule described as the“ condition to be satisfied by the diagnosis target ”” in the read rule (that is, , The pattern rule of the CHECK part of the rule) is described therein and is checked by recursively calling the "pattern check means" 501.

FIG. 6 shows an example of a code input by the code automatic diagnosis apparatus of the present invention, which is in C language.
This code example shows two types of processing for initializing an array of size a with 0 in C language. C
In the language, for example, when an array a of size 5 is declared as int a [5]; as follows, a is a [0], a [1], ..., A
The array has five elements [4] (up to a [4] instead of up to a [5]). The for sentence is a syntax that describes repetition. Here, in order to initialize all these elements to zero by the for statement, two fo shown in FIG. 6 are used.
How to write r sentences (12 to 14 lines and 16 to 18 lines)
There is. At this time, when declaring the array, since it is described as "a [5]", the end condition of the for statement is also "i <5".
It is considered easier to describe by using the size of the array as in. In other words, the description in lines 12 to 14 is easier to understand, and the description in lines 16 to 18 is harder to understand.

FIG. 7 shows an example of rules entered by the code automatic diagnosis device of the present invention. This rule relates to how to write a for statement that operates array elements in C language,
This is a coding rule that prohibits the writing of the for statement (lines 16 to 18) in the lower part of FIG. The rules are rule name, FIND section, PROC section, C
It is composed of an HECK section and an MSG section.

The first line is the rule name. The portions from the 2nd to 5th rows and the 8th and 9th rows are the FIND section, and specify the diagnostic target specific condition. The 6th and 7th lines and the 10th and 11th lines are the PROC part, and specify the processing of the diagnosis target information. The 12th and 13th lines are the CHECK part, which specifies the condition to be satisfied by the diagnosis target. 14, 1
The part of the fifth line is the MSG part and specifies the diagnostic message.

In the 3rd to 5th lines, the diagnostic target specific condition "for statement that repeats from 0 to integer constant N" is specified. In the 7th line, "processing of diagnostic target information" of "substituting the range of compound statement W corresponding to repeated description of for statement into a special variable called FIND range" is specified. The FIND range indicates the range of search in the next FIND section, and the default value (default value) is the entire code. In the 9th line, the diagnostic target specific condition "the array using the control variable: I: used for the repetition of the for statement as a subscript" is specified. The search range here is the range of the compound sentence W specified in the 7th line.

In the 11th line, "processing of diagnostic target information" of "substituting the size of the array A obtained in the 9th line into the variable M" is specified. In the 13th line, "the value obtained by subtracting 1 from the variable M does not match the integer constant N in the 3rd line".
"Conditions to be met by the diagnosis target" are specified. If they match, it is considered that the description should be prohibited, such as lines 16 to 18 in FIG.
Need to be denied. The 15th line specifies a diagnostic message that is output when the “condition to be satisfied by the diagnosis target” is not satisfied.

In the description of FIG. 7, the description sandwiched between "/" is a description representing a pattern rule (hereinafter referred to as a pattern description). -There are four types of descriptions that appear between rules, which are sandwiched by ":": one that represents a syntax element, one that represents a variable, one that represents a table, and one that represents another rule. Of these, those that represent syntax elements, variables, and tables are collectively referred to as an embedded description. -The part after ";" in the embedded description is the name given to the description. Once named, the embedded description can be referenced in the form ": name:". The function AREA () is a function for obtaining the range on the code of a specific pattern description. -The built-in description representing a syntax element is ": syntax element name :.
By writing in the form of "attribute name", the attribute of the syntax element can be referenced (Example: A: .size).% S,% d in the diagnostic message are: I:,:
It is a control character for editing and displaying M: as a character string and an integer, respectively.

FIG. 8 is an example of a diagnostic message output by the code automatic diagnostic apparatus of the present invention, which is a diagnostic message when the code shown in FIG. 6 is diagnosed based on the rule shown in FIG. The file name represents the file name of the code. The line number is the line number of the code and represents the matched position of the rule. The rule number represents the rule number of the checked rule. The message is a compilation of the MSG part of the rule.

FIG. 9 shows an example of the processing procedure of the rule control means 107 in the code automatic diagnosis apparatus of the present invention, which is the processing procedure in the case of the rule shown in FIG. The relationship between the configuration of the code automatic diagnosis device of FIGS. 1 to 5 and the code of FIG. 6, the rule of FIG. 7, and the diagnostic message of FIG. 8 will be described along the processing procedure of FIG.

First, the rule control means 107 reads the rule shown in FIG. Next, the rule control means 107 searches for a portion satisfying the diagnosis target specifying condition designated in the FIND portion of the rule (lines 2 to 5 in FIG. 7) (901, 90).
2, 903).

In step 903, the rule control means 107
Calls the "means for determining the diagnostic target specific condition" 108 to make this determination. At this time, "absolute position of the code"
Is also passed. Then, the first logic judgment means 201 in the “means for judging the diagnosis target specifying condition” 108 receives the diagnosis target specifying condition. Here, since the diagnosis target specifying condition is made up of only the pattern rule, the first logic judging means 201 passes this pattern rule to the pattern checking means 501. Pattern checking means 501
Determines whether the coding pattern of the code from the "absolute position of the code" matches the pattern rule.

If they do not match, the loop of 901 and 902 is performed while shifting the "absolute position of the code" column by column. When the “absolute position of the code” comes to the position of “f” of “for” on the 12th line of FIG. 6, the pattern rule and the 12th line of FIG. 6 start to match, but in the code of FIG. Since the end condition of is "i <5" instead of "i <= 5", it does not match.

Next, "absolute position of code" is 16 in FIG.
When it comes to the position of "f" (5th column) of "for" of the line, the pattern rule and the 16th to 18th lines of FIG. 6 match. Therefore, as the “pattern matching position”, the fifth column of 16 lines is written in the storage unit 109. The syntax elements in the pattern rule :: integer; I :: integer constant; N :: compound sentence; W: respectively match i 4 a [i] = 0; in the code (correctly, code information DB). This correspondence relationship (the correspondence relationship between the name given to the syntax element in the pattern rule and the information about the syntax element in the code information DB) is written in the storage means 109. (Hereinafter, referred to as “syntactic element correspondence information”.) The above “pattern matching position” and “syntax element correspondence information” are “specified diagnosis target information”.

Next, the rule control means 107 changes the "processing of diagnosis target information" designated in the PROC (6th and 7th lines in FIG. 7) of the rule to "processing means for processing diagnosis target information".
It is passed to 110 and the "processing of processing the diagnosis target information" 90
Do 4. In the "means for processing diagnosis target information" 110, the "processing distribution means" 301 calls the "range information extraction means" 302, and the processing in the seventh line of FIG. 7 is executed. Here, the range of the description “a [i] = 0;” on the 17th line of the code in FIG. 6 represented by: W: (9 on the 17th line)
(Column to column 17) is assigned to a special variable called "FIND range".

Next, the rule control means 107 searches for a portion satisfying the diagnosis target specifying condition designated in the FIND portion of the rule (8th and 9th lines in FIG. 7) (905, 90).
6, 907). The search range here is an area indicated by the special variable “FIND range” as indicated by 905 and 906.

The rule on the ninth line in FIG. 7 is a pattern rule as before. Here, the syntax element named: I: in the previous search, that is, the array using i as the subscript in this case, is searched. When referring to a syntax element that has previously appeared at the time of checking a pattern, the “syntax element description checking means” 509 refers to the “syntax element correspondence relationship information” written in the storage means 109. Match only those that have the same face as the ones that appeared previously. The syntax element in the pattern rule :: array; A: matches a in the code, and this “syntax element correspondence relationship information” is written to the storage unit 109.

Next, the rule control means 107 changes the "processing of diagnosis target information" designated in the PROC section (10th and 11th lines of FIG. 7) of the rule to "means for processing diagnosis target information". The processing is passed to 110, and the “processing of processing diagnosis target information” 908 is performed. "Means for processing diagnosis target information"
In 110, the “processing distribution means” 301 calls the DB operation means 303, and the processing of the 11th line in FIG. 7 is executed. Here, the DB operating means 303 causes the code information D
By searching B106, the attribute "size" of the array: A: (a in the code) (here, the value is 5) is obtained, and this is converted into an integer type variable named M. Substituted. Then, the information of the integer variable M is written in the storage unit 109. This is the diagnosis target processing information.

Next, the rule control means 107 is a means for checking the "condition to be satisfied by the diagnosis target" specified in the CHECK part of the rule (12th and 13th lines in FIG. 7). "111", and it is determined whether or not the "condition to be satisfied by the diagnosis target" is satisfied (909).
"Means for checking conditions to be satisfied by the diagnosis target" 11
In 1, the second logic judgment means 404 calls the “comparison condition check means” 402, and the check in parentheses on the 13th line in FIG. 7 is performed. Here,: N: is 4, and:
Since M: is 5, the expression in parentheses is "true". The NOT attached outside the parentheses is judged by the second logic judging means 404, and the “condition to be satisfied by the diagnosis target” is “false”.

Therefore, the rule control means 107 passes the diagnostic message specified in the MSG part of the rule (14th and 15th lines in FIG. 7) to the "diagnostic message output means" 112 and performs "diagnostic message output processing". Perform 910. The diagnostic message output in this way is shown in FIG.

After that, the rule control means 107
Returning to the judgment of 907 by the loops of 5 and 906, there is no array satisfying the condition of the 7th row of FIG.
Return to judgment of 03. Again, f that satisfies the determination of 903
Since there is no or statement, the rule check ends by exiting the loop of 901 and 902.

1 to 5, the rule control means 107 has a "means for judging a diagnosis target specific condition" 108, a "means for processing diagnosis target information" 110,
"Means for checking conditions to be satisfied by the diagnosis target" 11
By calling 1 and multiple stages, it is now possible to check rules that could not be checked in the past.

Next, the code automatic diagnosing device of the present invention executes the code (program in C language) shown in FIG.
An example of diagnosing based on the rules shown in 1 to 13 and outputting the diagnostic message shown in FIG. 14 will be described.

The rule of FIG. 11 is an example of using the "absolute position checking means" 504. As the diagnosis target specifying condition, a pattern rule including a special character that specifies an absolute position is specified in the FIND part. As a result, the pattern matching position is located at the 1st row and 1st column. Next, the pattern rule of the CHECK part is designated as the “condition to be satisfied by the diagnosis target”. This pattern rule expresses a standard comment format to be described from the pattern matching position. In the code of FIG. 10, the comment format as specified by the rule is described from the first row and the first column. Therefore, no diagnostic message is output. Also, in the pattern rules, "/" and "
Since "*" has a special meaning, the escape symbol "\" is used to express "\ /" and "\ *" to represent "/" and "*" as simple characters. ""Also"
"_" Also represents a blank. "_" Is used to facilitate counting the number of blanks.

The rule of FIG. 12 is an example of using "means for converting a plurality of lines of coding pattern rule into one line" 506, "relative position checking means" 505, and "paired description checking means" 507. is there.

This rule represents "if sentence standard pattern". First, the coding pattern "if and arbitrary number of blanks and left parenthesis" in the designation of the FIND section is searched. The escape symbol "\" before "(" is "("
The special meaning of "paired description" is negated, and "(" is expressed as a mere character. In the CHECK part, the standard pattern for describing the if sentence is specified. Explain the special characters included in the rules ・ "@" means "any number of blanks, optional comments,
It is a special character that represents the description "arbitrary number of blanks, line feed". By the way, when expressed by a regular expression, / _ * <\ / \ *. * \ * \ />? It becomes _ * \ n /. -"<", ">" Represent a block. ”<...
When expressed as "*", it means that the entire block sandwiched by "<" and ">" is repeated zero or more times. "" Indicates that the previous character or block may or may not be present. ． ． If "@" and "*" are described consecutively like "@ *", the description for one line before "@" is "
In the sense that "*" is added between <"and">"
<. ． ． This is equivalent to @> * ". In particular, this conversion is performed by" means for converting a multi-line coding pattern rule into one line "506 by adding a special character indicating a relative position to the multi-line coding pattern rule. To the
After this conversion, they are joined in one line. For example, the description "△△△△: Sentence: @ *" in the 6th and 9th lines is "
The special character that indicates the relative position of "△ * ¥ # △△△△: Sentence: @" is added to the one without "*", and then it is added to "<△ * \ # △△△△: Sentence. : @> * ”Is converted and then connected.

Therefore, since the pattern rule of the CHECK part is spread two-dimensionally, it is converted into the following one-line rule by the "means for converting a coding pattern rule of a plurality of lines into one line" (horizontal width). There is not enough, so it is written over multiple lines for convenience). / If △ (△: conditional expression: △) △ {@ ＜ △ ＊ ¥ # △△△
△: Sentence: @ ＞ ＊ △ ＊ ¥ #} @ △ ＊ ¥ # else △ {@ <
△ ＊ ¥ # △△△△: Sentence: @ ＞ ＊ △ ＊ ¥ #} @ /

"¥ #" in the above pattern rule is
Column position of "start position of pattern check"
It is a special character that indicates the relative position of the same column position as the first character "i" of the pattern rule. “¥ #” is checked by the “relative position checking means” 505.

When this rule is applied to the code of FIG. 10, the form of the if statement on lines 51 to 54 is grammatically correct, but it does not satisfy the rule (the executable statement at the time of conditional judgment of the if statement is "{ Since it is not surrounded by ","} ", a diagnostic message is output (the message on the first line in FIG. 14).
The form of the if statement on the 61st to 67th lines satisfies the rule. In order to satisfy the rule, it is useless if the pattern checked by the rule and the column position are deviated by one.

Also, "(", ")" in the pattern rule
Also, since "{", "}" is a "paired description", it matches only the paired description in the code. Therefore, even if the if statement has a multiple structure like the 71st to 87th lines, its own else clause and pattern are always checked without matching with the else clause of another if statement. . For example, in line 82, "}"
Since it is shifted by one column, a diagnostic message is output for the if statement starting from the 72nd line corresponding to this description (the message on the 2nd line in FIG. 14).

In this way, "pattern checking means"
By 501, a strict coding pattern rule can be checked. In addition, the coding pattern rule is not always strict. In the pattern rule, "△" is replaced with "△ *" (one blank is acceptable with any number of blanks) ・ "△△△ in indentation" By replacing “Δ” with “<ΔΔ> *” (the indentation of four columns is considered to be indentation of even columns), the strictness of the pattern rule can be freely relaxed.

The rule of FIG. 13 is an example of using the “reference means of another pattern rule” 502. In this rule,
In the pattern rule of the FIND section, another pattern rule "i
The "f sentence standard pattern" is referred to. This is the rule shown in Fig. 12. When the different pattern rule is referred to in this way, the "reference description of the different pattern rule" portion (in this case, Fig. 13). FIND part), the pattern rule of the CHECK part of the referenced rule (in this case, C of FIG. 12).
It operates as if the pattern rule of the HECK part) is written. This pattern rule check is performed by the “pattern rule control means” 502 calling the “different pattern rule reference means” 508.
“Reference means of another pattern rule” 508 is a rule DB
104 to search for another pattern rule (in this case, "if
When the sentence standard pattern ") is found, the pattern is checked by recursively calling the" pattern checking means "501 with the pattern rule of the CHECK part and the current" absolute position of the code "as parameters. In this way, when the diagnosis target is specified, next, CHECK in FIG.
The department is checked. Here, the special variable NESTLEVEL is the pattern matching position, in this case, "if
It represents the nesting level at the position where the "i" of the sentence standard pattern "matches. This is the" means for finding the nesting level "4
01. In the case of FIG. 10, the if statement starting on the 74th line does not satisfy the condition because the nesting level is 4, and the diagnostic message on the 3rd line of FIG. 14 is output.

In this way, it is possible to specify the absolute row and column position in the code and check the coding pattern to be described therein. Also, since the coding pattern rules spread two-dimensionally can be described as they are, coding pattern rules that are easy to write and easy to read can be described. In addition, the strict coding pattern rule can be checked by the "pattern check means" 501. In addition, the strictness of pattern rules can be relaxed freely. In addition, it is possible to check whether or not the corresponding description satisfies the coding pattern rule even in the coding pattern having the multiple structure by the “paired description checking means” 507. In other words, it is possible to avoid erroneous diagnosis in which an unnecessary diagnostic message is output by matching a description that does not correspond, and thus it is possible to save time for confirming the erroneous diagnosis. In addition, you can reuse another pattern rule,
Coding pattern rules can be expressed concisely and can be easily modified.

Next, the rules shown in FIGS. 15 and 16 that can be processed by the code automatic diagnosis device of the present invention will be described. FIG. 15 is a rule example showing that not only the coding pattern rule but also the naming convention of the user-defined word can be checked by the configuration of this embodiment. The FIND part is made up of a logical combination of pattern rules and "attribute conditions that the syntax element must satisfy". The “attribute condition to be satisfied by the syntax element” is checked by the “attribute condition check means to be satisfied by the syntax element” 202. CHEC
Part K has the shape of the comparison condition, but the right side has the shape of the pattern rule, which indicates that the pattern check is performed. In such a case, the “comparison condition check means” 402 calls the “pattern check means” 501 to perform the check. This rule checks the naming convention that pointer-type functions must start with "pf".

Further, with respect to the "specified diagnosis target information" and the information obtained by processing the information, for example, concatenation of character strings, extraction of partial character strings, or prefix (prefix) or postfix (suffix) By including a character string operation means for performing a character string operation such as removal of "in the means for processing the diagnosis target information" 110, a stronger naming convention can be checked.

FIG. 16 is an example of a rule showing that the definition order of the user-defined words on the code can be checked by the configuration of this embodiment. Here, the rule "the definition of the variable name should come after the function name is defined" is expressed. In the PROC section, first, a table A in which attributes and ranks are paired is created, and then the code information DB
The user word table in which the user defined words in 106 are stored in the order of definition is combined with the table A to create a table C in which the “character surface of the user defined word” and the “rank of the table A” are paired. The above processing is performed by the DB operating means 303. In the CHECK part, it is checked whether the second column of table C is in ascending order (A represents ascending order).
CHECKKORDER () is a function provided by "means for checking arrangement condition of table element" 403 and can check arrangement condition of table element.

Thus, the "pattern checking means" 5
01 can be used to check not only coding pattern rules but also user-defined word naming conventions. Also, DB
It is possible to check the definition order of the user-defined words on the code by using the operation means 303 and the "table element arrangement condition checking means" 403.

“Means for processing diagnosis target information” 110
However, for the "specified diagnosis target information" and the processed information, for example, an attribute such as converting a character string representing a numerical value into a numerical variable or conversely converting a numerical variable into a character string representing a numerical value. Attribute conversion means for performing conversion, numerical operation means for performing four arithmetic operations on the "specified diagnosis target information" and the numerical data in the processed information to process it into new data. "Specified diagnosis target information" or data in a table in which information is processed (of information in the code information DB that meets certain conditions and is arranged in the order of appearance in the code) For example, if a means such as a sort processing means for performing sort processing and processing into new data is included, the rule checking ability is further enhanced.

Further, "means for processing information to be diagnosed" 1
If the “means for designating the nesting level (depth of the hierarchical description) in which the diagnostic target should exist” included in 10 is also included in the “means for determining the diagnostic target specific condition” 108, the diagnostic target You can strengthen certain abilities.

If the code analysis means includes counting means for counting the number of lines of code and the number of calling functions included in the code, the rule checking ability is further enhanced.

[0090]

As described above, the code automatic diagnosis device according to the present invention has the following effects. (1) The ability to describe coding rules can be significantly improved. That is, in the conventional code automatic diagnosis device, a complicated coding rule, which cannot be checked without adding a check means for a specific rule to the device, can be checked without changing or expanding the device. As a result, it is possible to describe as a coding rule an expert's intuition or subjective (subtle or elaborate) empirical rule that cannot objectively or logically prove correctness. By expressing the coding know-how of such an experienced programmer as a coding rule, it is possible to teach a good coding style to a beginner who uses this automatic code diagnosis device.

(2) A strict check of the coding pattern can be performed. In other words, since it is possible to check up to the position of the row / column, it is possible to check a strict coding rule so that there is no difference in coding patterns between people. This makes it possible to thoroughly standardize coding patterns. (3) Since the absolute position can be specified in the pattern rule, the coding pattern rule can be expressed simply. (4) It is possible to specify an absolute line position where a coding pattern should be placed, which could not be conventionally diagnosed. (5) Since the two-dimensional spread of the coding pattern can be expressed as it is as the rule of the coding pattern, the rule is easy to write and easy to read. This makes it possible to prevent a description error or a correction error in the rule and facilitate comparison with the code to be diagnosed.

(6) Since the description of the pair of rules can be matched only with the description of the pair of codes, it is possible to express the rule as perceived by a human, prevent malfunction, and analyze many malfunctions required. You can save time. (7) Since the coding rule is expressed as a combination of some check processing and data processing, the content of the rule is expressed strictly and objectively. That is, the user can smoothly modify, add, or delete the rule without misunderstanding what is checked by the rule. (8) By reusing the rule once defined,
It has the following effects. That is, the description of a rule that appears many times can be made into a component to improve the description efficiency of the rule. Also, since the same description does not appear many times, you only have to modify one place when modifying, adding or deleting rules,
The maintainability of rules increases. Furthermore, the user can easily write more sophisticated rules by combining the rules defined so far.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of an automatic code diagnosis device of the present invention.

FIG. 2 is a diagram showing an internal configuration of a “means for determining a diagnosis target specifying condition” in FIG.

FIG. 3 is a diagram showing an internal configuration of a “means for processing diagnosis object information” in FIG. 1.

FIG. 4 is a diagram showing an internal configuration of “means for checking conditions to be satisfied by a diagnosis target” in FIG. 1.

FIG. 5 is a diagram showing an internal configuration of a “pattern check unit” shown in FIGS. 2 and 4;

FIG. 6 is a diagram showing a first example of a code input by the automatic code diagnosis device of the present invention.

FIG. 7 is a diagram showing a first example of rules entered by the code automatic diagnosis device of the present invention.

FIG. 8 is a diagram showing a first example of a diagnostic message output by the code automatic diagnostic device of the present invention.

FIG. 9 is a diagram showing an example of a processing procedure of rule control means in the code automatic diagnosis device of the present invention.

FIG. 10 is a diagram showing a second example of a code input by the automatic code diagnosis device of the present invention.

FIG. 11 is a diagram showing a second example of a rule input by the code automatic diagnosis device of the present invention.

FIG. 12 is a diagram showing a third example of a rule input by the code automatic diagnosis device of the present invention.

FIG. 13 is a diagram showing a fourth example of a rule input by the code automatic diagnosis device of the present invention.

FIG. 14 is a diagram showing a second example of a diagnostic message output by the code automatic diagnostic device of the present invention.

FIG. 15 is a diagram showing a fifth example of a rule input by the code automatic diagnosis device of the present invention.

FIG. 16 is a diagram showing a sixth example of rules input by the code automatic diagnosis device of the present invention.

FIG. 17 is a schematic block diagram showing an embodiment in which the code automatic diagnosis device of the present invention is configured as a microcomputer system.

[Explanation of Codes] 101 ... Rule 102 ... Code 103 ... Rule Input Means 104 ... Rule DB 105 ... Code Analysis Means 106 ... Code Information DB 107 ... Rule Control Means 108 ... Means for Determining Diagnostic Target Specific Condition 109 ... Storage Means 110 ... means for processing information to be diagnosed 111 ... means for checking conditions to be satisfied by the diagnosis object 112 ... means for outputting diagnostic message 113 ... diagnostic message 201 ... first logical determination means 202 ... check for attribute condition to be satisfied by syntax element Means 301 ... Process distribution means 302 ... Range information extraction means 303 ... DB operation means 401 ... Nest level finding means 402 ... Comparison condition checking means 403 ... Table element array condition checking means 404 ... Second logical judgment Means 501 ... Pattern checking means 02 ... Pattern rule control means 503 ... One character matching check means 504 ... Absolute position check means 505 ... Relative position check means 506 ... Multiple-line coding pattern rule conversion means 507 ... Paired description Checking means 508 ... Reference means for another pattern rule 509 ... Checking means for syntax element description 601 ... CPU 602 ... Main storage device 603 ... Input / output terminal device 604 ... Printer output device 605 ... External storage device

Claims (6)

[Claims] 1. A device for automatically diagnosing a code (102) coded according to a predetermined grammar and input to a computer based on a separately defined coding rule, the device comprising: Enter the rule (101) expressed in a format that can be interpreted by the device and write the rule information to the rule DB
Rule input means (103) for storing in (104) and code analyzing means (105) for analyzing the code (102) and storing code information in the code information DB (106).
And a "means for determining a diagnostic target specific condition" for determining whether a specific position of the code (102) satisfies a diagnostic target specific condition that is a condition for specifying a diagnostic target specified in the rule information ( 108) and the "processing of diagnosis target information" specified in the rule information.
"Means for processing information to be diagnosed" for executing (110)
And "means for checking conditions to be satisfied by the diagnosis target" (111) specified in the rule information, and "determining the diagnosis target specific condition" according to the rule information. Rule control means (107) for independently calling "means for performing" (108), "means for processing information to be diagnosed" (110) and "means for checking conditions to be satisfied by the diagnosis target" (111), Means for determining the diagnostic target specific condition "(108),
"Means for processing information to be diagnosed" (110) and "Means for checking conditions to be met by diagnosis object" (111)
A storage unit (109) for writing and referring to information in the processing process, and called by the rule control unit (107) when the "condition to be met by the diagnosis target" is not satisfied,
An automatic code diagnostic device based on a coding rule, comprising: "diagnostic message output means" (112) for outputting a diagnostic message (113) according to the rule information. 2. A device for automatically diagnosing a code (102) coded according to a predetermined grammar and input to a computer based on a separately defined coding rule, the device comprising a rule for a code and a coding pattern. A "pattern checking means" (50) for checking whether or not the code (102) satisfies the rule for the coding pattern by comparing and matching one by one character.
1) An automatic code diagnosing device based on a coding rule, including: 3. The "pattern checking means" (50
1) is "absolute position checking means" (50) for checking whether or not a character or description included in a rule of a coding pattern is placed at a certain absolute position on the code.
4), The code automatic diagnosis device based on the coding rule according to claim 2. 4. The “pattern checking means” (50
In 1), the coding pattern rule that spreads two-dimensionally is decomposed line by line, and the pattern rule is checked while correcting the relative position from the previous line. Means "(506,
505), The code automatic diagnosis device based on the coding rule according to claim 2. 5. The "pattern checking means" (50
1) is called when there is a paired description in the pattern rule, and checks that the paired description on the code and the paired description on the rule correspond to each other. Automatic checking device based on the coding rule according to claim 2, further comprising: "checking means" (507). 6. The “pattern checking means” (50
1) is called when there is a "reference description of another pattern rule" in the pattern rule, fetches another pattern rule from the rule DB (104), and checks the pattern according to the other pattern rule. The code automatic diagnosis device based on the coding rule according to claim 2, further comprising: "a means for referring to another pattern rule" (508) for executing.