JP2005346409A - Program correction support system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To overcome the drawback that there is no method for automatically detecting failure regarding check of such correction spots aimed at preventive maintenance in relation to a method and a system for supporting correction of a program, resulting in omission in the check, or inconsistency is resulted because the correction policy is not directly conveyed to developers and thus the correction is not synchronized. <P>SOLUTION: The defect is overcome by a program correction support system which realizes a correction instruction on a top-down basis by the constitution of an overall management server 1 for issuing the correction instruction by distributing the management of source code groups 36, 46 by client unit to make source management such as keyword retrieval to be performed automatically; a product management server 2 which relays the correction instruction and integrates results in an influenced range by product, and a client 3 which automatically extracts source code corresponding to the correction instruction. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a method and system for supporting program correction.

Conventionally, when a problem is found in a library function or how to use the function, as a preventive maintenance of the program, the program developer himself finds a part of the program that he / she uses and uses the function in question. The program has been modified in accordance with its own policy. For large programs, the amount of code in the program is large and there are many developers, so file search is used to find a specific part, and a consistent correction policy is communicated to the developer by means such as e-mail. , We have tried to unify the correction of problem areas. In program change management, as disclosed in Japanese Patent Application Laid-Open No. 2002-182916, there is one that centrally manages source modules and extracts an influence range of a specific modification.

JP 2002-182916 A

The above prior art only prompts related developers to confirm specific program changes without omission, and is not configured to present similar corrections in multiple projects and multiple products. . In addition, since the developer is the input of the program change, the controlled instruction system from the administrator to the developer cannot be established. Instructions must be given by means of communication. Since the change confirmation and the correction policy are not synchronized, there is a problem that confusion occurs when the correction policy is changed at a short interval with respect to the same correction portion.

The object of the present invention is to manage the source code of a program and automatically extract correction points for each developer machine in response to top-down correction instructions. The main feature is to provide a program modification support system that does not miss the modification time by directly communicating with developers.

In order to achieve the above object, a program developed on a developer's client is mechanically managed on the client to provide a program code checking function. Furthermore, it has a management server that issues a code correction check instruction as a message, and has a function of executing a code correction check according to the message and returning the check result to the client and the management server. Furthermore, it has a function to put a correction policy in the message and display the correction policy on the client machine together with the check result.

According to the present invention, since the source code developed on the developer's client machine can be automatically and mechanically checked by the check function, in the source code created by the developer by the client, the omission of checking the corrected portion is eliminated. There is an advantage that can be. In another embodiment of the present invention, the management server is placed on a separate machine, and correction check instructions and policies can be simultaneously deployed to a plurality of clients. Even when the revision policy changes, there is an advantage that the revision policy can be transmitted in a timely manner.

  The best mode for carrying out the present invention will be described below in detail with reference to the drawings.

FIG. 1 shows a program correction support system according to an embodiment of the present invention. In the present embodiment, the overall management server 1 and product management server 2 connected to the network 5.
, The client 3, and the source server 4, and the general management server 1, the product management server 2, and the client 3 respectively include an attribute information management unit (11, 21, 31) and a correction message creation unit (12, 22). 32), a message processing unit (13, 23, 33), a result reporting unit (14, 24, 34), and a source code management unit (15, 25, 35).

The overall management server 1 is a management server serving as a route for managing corrections, and there is only one, which manages the source codes of all products. The product management server 2 is a server that manages the source code of a specific product among the products handled by the company, and there are a plurality of products in a company that handles a plurality of products. The client 3 is a terminal used by a developer engaged in product development, and manages the source code developed by the client 3. In this example, overall management,
Product management and client have three layers, but for products with a large development scale, the component management server that manages the source code of a specific component among the components included in the product is connected between the product management server and the client. It is also possible to have four layers. That is, the product management server is configured to occur under the management of a plurality of component management servers, and the client is placed under the management of the component management server.

Since the overall management server 1 does not perform direct development, the source code management unit 15 does not function. If development is not performed in the product management server 2 as well, the source code management unit 2
5 does not function, and if no correction check instruction is issued, the correction instruction message creation unit 22 may not function. Since the client 3 performs actual development, the source code is managed. The source code group to be managed may be in the client machine (source code group 36) or in another source server 4 ( Source code group 46) It is only necessary to be in a location (path) that can be referred from the client machine and can confirm the existence of the source code file. Since the client 3 is located at the end and does not monitor another machine, the client 3 does not issue a correction instruction, and the correction instruction message creation unit 32 does not function.

FIG. 2 shows an outline of the flow in the present invention from when a correction instruction is issued until the administrator grasps how much the influence of the correction instruction is. First, the administrator inputs a correction instruction and a correction policy from the correction instruction message creation unit of the overall management server (101). The correction instruction message is processed by the message processing unit. If the role of the machine is not “client” in the attribute information, the correction instruction message is transmitted to the managing client. The message is transferred as it is to each product management server if it is a general management server, and to each client if it is a product management server (102, 103). The client searches the source code (file) to be corrected from the managed source code according to the correction instruction message (104), lists the search results, creates a new message as a result message, It is passed to the message processing unit (105). If it is a result message, it is passed to the result report unit. If it is a client, a list of correction parts and a correction policy as a result are displayed in a pop-up on the client machine, and a correction instruction is issued to the developer (106). . Regardless of the role, the result is output as a log on each machine (107, 108, 109). By looking at the result log, the product management server can grasp the range of influence of the correction on the product being handled, and the overall management server can grasp the range of influence of the correction on all products. As an extension of the present invention, it is conceivable to link the result reporting unit with the pending management system. It is possible to eliminate omissions by making a list of corrections pending.

FIG. 3 shows the details of each configuration of the program correction support system of the present invention. The attribute information management unit 210 includes an attribute information setting function 211 for setting attribute information, and an attribute information providing function 212 for extracting and providing the set attribute information. The correction instruction message creating unit 220 includes a correction message input function 221 for an administrator to input a correction instruction message, and a correction message transmission function 222 for instructing transmission of a correction message.
It is included. The message processing unit 230 parsed and received the received message analysis function 231 that analyzes the received message and decides to send the message to an appropriate function or machine, and the message relay function 232 that sends the message to another machine. A transmission / reception check function 233 for managing messages and managing whether transmission / reception is properly performed is included. The result report unit 240 has a result log output function 241 for outputting a result message as a log,
Result display function 242 for displaying results on the client along with the correction policy
It is included. The source code management unit 250 includes a source code registration function 251 for registering the source code in the source code DB so that the source code is managed, and a source code search for searching the corresponding source code by operating the managed source code. A function 252, a result message creation function 253 for converting a search result into a result message, a result message transmission function 254 for instructing transmission of the result message, and a message for monitoring the existence of the source code are generated and transmitted. And a source code DB 256 in which source code information is managed.

FIG. 4 shows details of each function of the attribute information management unit 210. Attribute information setting function 2
11 has a GUI and sets machine attribute information in the attribute information file 330 on an input screen such as the attribute information setting dialog 319. The attribute information file manages three attributes. The role attribute 331 indicates the role of the machine. The management structure of the machine has a tree structure with the general management server as the root, and the role also depends on each part of the tree structure: “server” corresponding to the root, “relay” corresponding to the branching node, and leaves There are three types of corresponding “clients”. At least one of them is set. As a stand-alone,
It is possible to have two roles of “server” and “client”. The management client attribute 332 is an attribute indicating a client group managed by the machine, and describes a product management server group managed by the overall management server and a client group managed by the product management server ( 391, 392). For this reason, a client that does not have a machine to be managed does not have a management client attribute (393). The affiliation server attribute 333 is an attribute for indicating which machine is managed by the machine. The product management server describes the overall management server, and the client describes the product management server (392, 393). The general management server that is not managed by any machine has no attribute of the affiliated server (391). The attribute information providing function 212 is a function for indicating the contents of the attribute information, and is used by the message processing unit to obtain machine information of a partner to which the message is transmitted.

FIG. 5 shows details of each function of the correction instruction message creating unit 220. The correction message input function 221 has a GUI, and inputs a condition for specifying a target file to be corrected and a correction policy comment on an input screen such as a correction message creation dialog 419, and generates a correction instruction message. The correction instruction message transmission function 222 is a function for transmitting the message created by the correction message input function 221 to the message processing unit.

FIG. 6 shows details of the message 540 that circulates in the message processing unit 230 and the system in a specific format. The received message analysis function 231 is a function of analyzing the content of the received message and determining a function or machine to which a message is to be sent next. The message relay function 232 is a function for transmitting and receiving messages. In addition to a function for communication, the message relay function 232 encrypts and decrypts messages flowing through the network. The transmission / reception check function 233 is a function for checking message transmission / reception with another machine. In addition to a function for retransmitting when transmission to the destination machine fails, the transmission / reception check with the destination machine periodically. And check for missing messages.

A message 540 indicates a message distributed in the correction support system of the present invention. The instruction condition 541 indicates instruction contents, and describes conditions for specifying a search target file in addition to processing contents such as log output and search. The comment 542 shows contents supplementing the instruction. The instruction condition syntax 549 indicates the format of the contents described in the instruction condition 541. In the simple example of the present invention, three of file, key, and proc are used.
It is a syntax to set each content. There are several types of messages. The correction instruction message 591 is a “C ++” file as an instruction condition, and “funcX”.
"Search" is indicated for the source code that contains the keyword, and the revision policy is indicated in the comment. The check result message 592 only includes an instruction “log output” as an instruction condition, and the “result contents” to be output is included in the comment. The message has a unique message ID, and as indicated by 592, the result always includes the message ID of the instruction message. The monitoring message 593 is issued periodically by the source code management unit in order to monitor the source code managed by the client. The instruction condition indicates whether the “all” file is “found” or not. The check indicates that the output comment is not found in the comment.

FIG. 7 shows each function of the result report unit 240 in detail. Result log output function 241
Outputs a list of check results as a log such as log output 691. Not only the result of the correction instruction but also the file existence monitoring is output as a log. The result display function 242
It has a GUI and displays the check result on the machine on a screen such as a result display dialog 629. The check result is displayed in the check target 621, and a supplementary comment such as a correction policy is displayed in the comment. In the check result display 692 of the correction instruction, the check target displays a list of line numbers including the source code and keywords to be corrected, and the comment indicates the correction policy. In the result message 592, the correction policy is not included in the comment, but the correction policy is included in the “result contents”.
Displayed from D. In the check result display 693 of the file existence monitoring, a list of file names for which the existence of the file as a check target could not be confirmed is displayed, and the comment indicates that the file could not be confirmed.

FIG. 8 shows each function of the source code management unit 250 in detail. The source code management unit 250 manages the registered source code in the DB, and searches for the source code according to a message instruction. The source code registration function 251 is a function for registering the source code in the DB, and has a GUI such as the source code registration dialog 719 and inputs the location where the source code is stored. , And a field for inputting a source code language type (713), a source code creator (711), and a product name (714). The registered source code information is managed by the source code DB 256. The source code DB 769 is an example of the source code DB, and the table 79
It is managed by a DB table such as 0. One record (791) has information such as 799. When the source code search function 252 scans the source code DB 256 according to the message and extracts a corresponding file, the source code search function 252 checks whether there is a line containing the target keyword in the file and lists the corresponding ones. Therefore, if the file condition of the instruction message is “C ++”, “language” in the source code DB 769 is “cpp”.
Are extracted (793), the contents of the file in the “location” of the record are searched, and if there is a line in which a keyword such as “funcX” is embedded, the file name and line number are left as a result. . Whether a file extracted by scanning the source code DB exists is also checked at the same time. If it does not exist, it is left in the result with a message that the file cannot be found instead of the line number. The search result is converted into a message by the result message creation function 253. In addition to the search results, the message ID of the correction instruction message is included, and the message is formatted as a message. Result message transmission function 25
In 4, the created result message is sent to the message processing unit. The monitoring message transmission function 255 is for monitoring whether the source code exists, and automatically and regularly issues a message for monitoring.

The message processing unit can expand the contents to be processed, but in this embodiment, only three simple processes are performed. FIG. 9 shows the flow when the message transmission destination is determined by the received message analysis function of the message processing unit. The analysis of the message checks what processing the instruction content described in the message is (810, 820, 830).
If the instruction content is “log output”, it is sent to the result report unit (811), and if the role is “client” (812), it is sent again to the result report unit and the result is displayed on the screen by the result display function ( 813). If the role is not “server”, a message is also sent to the higher-level affiliated server (815). In the case of a search for “modification instruction” as the instruction content, if the role is “client” (821), it is sent to the source code management unit (822), the search is performed, and the role is “server”. If it is “relay” (823), a message is sent to the managed client (824). When the instruction content is “file monitoring”, if the role is “client” (831), it is sent to the source code management unit (832), and existence is checked. File monitoring is a special case of source code search. Since all source codes are searched with non-existing keywords, the existence of the file itself is checked, but no keyword search is performed. The keyword that does not exist is assumed to be a keyword larger than the size of the managed source code, and even if the source code exists, the file size is less than the keyword length, so the keyword search itself is performed. Absent.

It is explanatory drawing which showed the whole structure of the implementation method of this invention. Example 1 It is the flowchart which showed the flow of one correction instruction | indication. It is a figure explaining the function for every component with which a correction assistance system is provided. It is a figure explaining the function of the attribute information management part. It is the figure which corrected and demonstrated the function of the message preparation part. It is a figure explaining the function of the message processing part. It is a figure explaining the function of a result report part. It is a figure explaining the function of the source code management part. It is a flowchart which determines the transmission destination of a message.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Overall management server (server), 2 ... Product management server (relay), 3 ... Client (client), 4 ... Source server, 5 ... Network, 11 ... Attribute information management part (server)
, 12 ... correction instruction message creation unit (server), 13 ... message processing unit (server), 1
4 ... Result report section (server), 15 ... Source code management section (server), 21 ... Attribute information management section (relay), 22 ... Correction instruction message creation section (relay), 23 ... Message processing section (relay), 24 ... result report part (relay), 25 ... source code official part (relay), 31 ... attribute information management part (client), 32 ... modification instruction message creation part (client), 33 ... message processing part (client), 34 ... Result report section (client), 35... Source code management section (client), 36... Source code group (local), 46... Source code group (separate machine).

Claims (3)

An attribute information management unit that manages information indicating the role of the own machine, client information managed by the own machine as attribute information, a correction instruction message generation unit that inputs a correction instruction from the user and generates a correction instruction message; A message processing unit for transmitting a message generated in the system including the correction instruction message to a transmission destination determined based on the role of the attribute information, and a source code group, and designated by the correction instruction message A program correction support system comprising a source code management unit for creating a search result obtained by extracting a source code as a result message. 2. The program correction support system according to claim 1, further comprising a result report unit having a result log output function for outputting the content of the result message in a log and a result display function for displaying the content of the result message in a dialog. Program correction support system. 3. The program correction support system according to claim 2, wherein a correction policy can be input as a comment to the correction message creation function, and the correction instruction message including the comment can be transmitted and received as a correction instruction message with a correction policy. A program correction support system comprising: the message processing unit; and the result report unit having the correction result display function for displaying the comment of the correction instruction message with the correction policy together with the result message.

Images