WO2008009629A2

WO2008009629A2 - Method for producing a consistent state of a program component quantity of a computer program

Info

Publication number: WO2008009629A2
Application number: PCT/EP2007/057227
Authority: WO
Inventors: Christian Schaefer; Sadi Yigit; Ulf Dreyer
Original assignee: Robert Bosch Gmbh
Priority date: 2006-07-18
Filing date: 2007-07-13
Publication date: 2008-01-24
Also published as: DE102006033143A1; WO2008009629A3

Abstract

The invention relates to a method for producing a consistent state of a program component quantity (2, 2-1, 2-2, 2-3, 2-4) of a computer program, said method using a program component quantity (2, 2-1, 2-2, 2-3, 2-4) containing information (4, 4-1, 4-2) about the model forming the basis of the computer program, a source code (6, 6-1, 6-2), and a structure description (8, 8-1, 8-2). In a consistent state of the program component quantity (2, 2-1, 2-4), the source code (6, 6-1, 6-2) can be automatically generated from the model information (4, 4-1, 4-2) according to the structure description (8, 8-1, 8-2). Said method contains the following steps: a new source code (6-2) which contains variations in relation to an initial output source code (6-1) of a consistent initial program component quantity (2-1) is provided; at least some of the components (4-1, 8-1) of the initial program component quantity (2-1) are compared with the new source code (6-2); and a new structure description (8-2) and new model information (4-2) are produced on the basis of the comparison such that the new source code (6-2) can be automatically generated from the new model information (4-2) according to the new structure description (8-2).

Description

description

title

A method for establishing a consistent state of a program component set of a computer program

The invention relates generally to the field of automatically generating source code of computer programs. In particular, the invention relates to a method for establishing a consistent state of a program component set of a computer program.

State of the art

DE 10 2004 043 788 A1 relates to a device for creating executable programs. The device includes a user interface for inputting a model underlying a program and a generator for generating individual program code from the model.

DE 198 37 871 C2 relates to a method for automatically generating a status-based program for a component of a system consisting of a plurality of components communicating with one another. The specification of the system serves as an input to a computer which produces as output a state-based program for each of the components.

According to the prior art, it is customary for the generator or the computer, which automatically generates the program code or the program, to operate according to a predefined structure description, ie the manner in which the model or the specification of the program code is firmly implemented in the generator. As program components are stored here only the model information and the code. Another example is the Matlab-Simulink software, in which code data is generated from domain data and no combined processing of these two representations takes place. In practice, it is also known to conserve changes made directly at precisely specified locations in the source code and to reinstate them to the specified location in the event of a new automatic program generation.

In addition to code generation forward programming, refactoring tools, such as the Eclipse development environment (www.eclipse.org), are known, but they use only the code to make syntactic changes.

Disclosure of the invention

It is an object of the present invention to ensure that any changes that have been made directly in a source code automatically generated from model information and a structure description are also preserved in a subsequent change of the model information and the subsequent automatic generation of the source code.

This object is achieved by a method according to claim 1.

An essential point of the invention is that in addition to the model information and the source code and the structure description is stored and / or processed, according to which the source code can be generated automatically from the model information. In this sense, the model information, the structure description and the source code are processed and / or stored together.

An essential point of the inventive method is that using model information and structure description manual changes to the program, which are performed by a user / programmer, recognized, classified and returned to model and structure description.

A program component quantity according to the invention consists of or comprises the source code as well as a set of components to be stored which are required for the automatic generation of the source code. That is, the program component set contains at least the model information, the structure description, and the source code.

By integrating the structure description into the program component set, it is possible to use model information and the structure description To recognize changes in the source code, to classify them and to return them to the model information and the structure description. In particular, the method according to the invention provides a methodology for the detection and classification of changes in the automatic creation of programs.

Advantageous developments of the method according to the invention are specified in the subclaims 2 to 7.

By additionally storing and using the structure description, the source code and the model information, such as domain data, can be better correlated. This allows, for example, consistent changes to the models.

By returning source code changes to the model information and the structural description, it is ensured that the changes made to the source code are preserved, even if the model information is subsequently changed and the source code is automatically regenerated.

In contrast to known techniques, the enrichment of the model information and structure description can not be done only by dedicated tools. The user / programmer can enrich the model levels by normal programming (source code manipulation). In addition, changes that are not (yet) provided by the model are not lost in a new generation step and are always inserted in the correct place.

The above statements and definitions also apply to the following description of the drawings and vice versa.

Brief description of the drawings

The invention will be explained in more detail below with reference to an embodiment. It shows:

FIG. 1 schematically shows changes of state of a program component quantity when carrying out an embodiment of a method according to the invention; FIG. 2 shows an embodiment of a program component quantity according to FIG

Invention, and

FIG. 3 shows a further embodiment of a program component quantity according to the invention.

Embodiments of the invention

A program component quantity 2 according to the invention of a computer program contains model information 4 of the model underlying the computer program.

The model information 4 may be, for example, domain data. A domain refers here to a specific field of application, for. B. a programmable logic controller (PLC). Domain data is variable data that can be used to define a project in combination with knowledge of the domain.

A computer program may be, for example, a programmable logic controller (PLC) program, a user interface (HMI) or drive software (NC), or any other computer program.

The program component quantity 2 also contains a source code 6 which can be generated automatically from the model information 4 in accordance with a structure description 8. According to one embodiment of the invention, the source code 6 is directly executable. According to a further embodiment, the source code 6 can be compiled into an executable program. According to yet another embodiment, the

Source code 6 to be interpreted by an interpreter.

According to the invention, the program component quantity 2 also contains the structural description 8 according to which the source code 6 can be automatically generated from the model information 4. That is, the structure description 8 indicates how to generate the source code 6 from the model information 4. In this sense, the structure description 8 is linked to the model information 4. According to the invention, the structure description 8 as well as the model information 4 may be changeable; In particular, the structural description 8 may contain variable data.

The structural description 8 can be constructed in the simplest case similar to a void text, wherein the gaps are filled with the model information 4. Furthermore, the For example, structure description 8 specify with which methods the model information 4 is processed.

While in the prior art the model information 4 must comply with rigid specifications of the used code generator, according to the invention the three components 'model information 4', 'structure description 8' and 'source code 6' preferably represent equivalent components of the program component set 2 which together represent the computer program concerned. In particular, the three components are preferably stored together or processed together.

The three components 'model information 4', 'structure description 8' and 'source code 6' of the program component set 2 are preferably designed such that the third component can be derived from in each case two of the components.

The method according to the invention relates to establishing a consistent state of the program component quantity 2 of a computer program during manual changes in the source code 6.

In the prior art, manual changes in the source code 6 are lost because they are not returned to the model information 4.

FIG. 1 shows the chronological sequence of the states of the program component quantity 2 when carrying out an embodiment of the method according to the invention.

The starting point for the method according to the invention forms a consistent starting point.

Program component set 2-1 including output model information 4-1, an output structure description 8-1 and an output source code 6-1 as indicated by 'A' in FIG. Since the output program component quantity 2-1 is in a consistent state, the output source code 6-1 is producible from the output model information 4-1 according to the output structure description 8-1. The arrows 9 in Figure 1 indicate the direction of the sequence of states.

If changes are now made in the output source code 6-1 and thus a new source code 6-2 is generated, then the resulting output program component quantity 2-2 is in an inconsistent state, ie the new source code 6 can not be derived from the source code. Model information 6-1 and the output structure description 8-1 generate. This condition is indicated in Figure 1 with 'B'. In order to restore the program component quantity 2-2 to a consistent state, the method according to the invention comprises the following steps:

Comparing at least part of the output program component set 2-1 with the new source code 6;

and generating a new structure description 8-2 and new model information 4-2 based on the comparison of at least a portion of the source program component set 2-1 with the new source code 6-2, the new source code 6-2 being extracted from the new model information. 2 according to the new structure description 8-2 is automatically generated.

The thus created (new) program component set 2-4 with the new model information 4-2, the new structure description 8-2 and the new source code 6-2 is again in a consistent state. That is, subsequent changes in the new model information 4-2 automatically take into account the previous changes in the source source code 6-1. It is very advantageous here that after the consistent state of the program component quantity 2 has been established, no source code parts have to be taken into account if changes in the model, ie. H. be made in the model information 4.

According to a preferred embodiment of the invention, the comparison of the new source code 6-2 with at least part of the output program component set 2-1 is a comparison of the new source code 6-2 with the output model information 4-1 and the output structure description 8-1.

For comparison of the new source code 6-2 with the output model information 4-1 and the output structure description 8-1, a first tree structure can be constructed from the new source code 6-2, this first tree structure being based on a

Model information 4-1 and the output structure description 8-1 corresponding second tree structure is displayed. The construction of the first tree structure from the new source code 6-2 can be done with the usual compiler techniques.

Preferably, the generation of the new structure description 8-2 and the new model information 4-2 is performed automatically using the output model information 4-1, the output structure description 8-1, and the new source code 6-2. In constellations in which the new source code 6-2 can not be unequivocally attributed to the new model information 4-2 and the new structure description 8-2, it is provided that a user query takes place, and thus the generation of the new structure description 8-2 and the new model information 4-2 is performed using the output model information 4-1, the output structure description 8-1, the new source code 6-2, and a user interaction.

The generation of the new structure description 8-2 and the new model information A-1 can be done in two steps:

First, generate the new structure description 8-2 based on a first comparison of the at least a portion of the components of the source program component set 2-1 with the new source code 6-2. This is illustrated in FIG. 1 as an example in the state 'C'. The comparison is made here by comparing the output model information 4-1 and the output structure description 8-1 with the new source code 6-2, d. H. the at least part of the components of the output program component set 2-1 in the illustrated embodiment consists of the output model information 4-1 and the output structure description 8-1. In the embodiment shown in FIG. 1, only the new source code 6-2 for determining the new structure description 8-2 is used on new components (arrow 13).

Second, generating the new model information 4-2 based on a second comparison of a second part of the source program component set 2-1 with the new source code 6-2 and the new texture descriptor 8-2, wherein the new model information 4-2 does not structurally represents representable differences between the output program component set 2-1 and the new source code 6-2. This is shown in Figure 1 in the state 'D'. In the illustrated embodiment, the second part of the output program component set 2-1 consists of the output model information 4-1. In the embodiment shown in Figure 1 is on new

Components of the new source code 6-2 (arrow 15) and the new structure description 8-2 (arrow 17) used to determine the new model information 4-2.

As a result, the inventive method again leads to a consistent state of the new program component set 2-4, which includes the new model information 4-2, the new structure description 8-2 and the new source code 6-2. The timing according to the illustrated embodiment of the method is 'A'->'B'-> 'C ->' D '. The new program component quantity 2-4 can itself be output program component quantity for a renewed execution of the method according to the invention, which is symbolized in FIG. 1 by the arrow 11 between state 'D' and state 'A'. Furthermore, the new program component set 2-4 may be the starting point for a change in the new model information 4-2 and / or in the structural description 8-2, with the changes previously made in the source code 6-1 (leading to the new source code 6-2 ), since these changes in the source code were raised according to the invention to the model level 4 and the structural level 8.

After generating one of the new structure description 8-2 and the new model information 4-2, the new source code 6-2 can be automatically generated from the new model information 4-2 according to the new structure description 8-2.

The model information 4, the structure description 8 and the source code 6 can be stored in a single file.

In particular, the model information 4, the structural description 8 and the source code 6 can be stored in different sub-trees 10 of a single XML file 12, as shown in FIG.

Alternatively, the model information 4, the structure description 8 and the source code 6 may each be stored in a separate file 14, as shown in FIG.

The file 12 or the files 14 may be stored in a conventional manner on a storage medium 16, for example a contiguous storage medium such as a CD (Figure 2) or on a distributed storage medium such as a network (Figure 3). In particular, the invention includes a storage medium 16 having at least one program component set 2 stored thereon. The illustration in the drawings is merely exemplary here. Any other types of storage on storage media are possible.

A method of storing a program component set 2 according to the invention may include the step of storing the texture description 8, and may further include the step of storing the model information and / or including the step of storing the source code. The processing of a program component quantity 2 according to the invention can be carried out according to a method which comprises the step of recording the structure description 8 as a variable input quantity. The program may further comprise the inclusion of at least one other of the remaining components of the program component set 2, in particular 'model information 4' and 'source code 6'.

A method for processing a program component quantity 2 of a computer program according to the invention can, for example, be a method for automatically generating a source code of the computer program, the method in this case comprising the following steps:

Taking model information 4 of the model underlying the computer program as a variable input variable;

Recording the structure description 8 as a variable input size; and

automatically generating the source code 6 from the model information 4 according to the structural description 8 by means of a code generator.

'Recording' in this sense can be for example a 'reading in from a storage medium'.

The invention is not limited to the embodiments shown in the drawings. Rather, it results from a general expert consideration of the description, the drawings and the claims.

Claims

claims

1. A method for establishing a consistent state of a program component set (2, 2-1, 2-2, 2-3, 2-4) of a computer program, wherein the program component set (2, 2-1, 2-2) contains model information ( 4, 4-1, 4-2) of the model underlying the computer program, a source code (6, 6-1, 6-2) and a

Structure description (8, 8-1, 8-2) contains and in a consistent state of the program component set (2, 2-1, 2-4) of the source code (6, 6-1, 6-2) from the model information (4, 4-1, 4-2) according to the structural description (8, 8-1, 8-2) is automatically generated, and wherein the method comprises the following steps:

Providing a new source code (6-2) having changes to an output source code (6-1) of a consistent output program component set (2-1);

Performing a comparison of at least a part of the components of

Output program component set (2-1) with the new source code (6-2), and

Generating a new structure description (8-2) and new model information (4-2) based on this comparison such that the new source code (6-2) is obtained from the new model information (4-2) according to the new structure description (8-2) is automatically generated.

2. The method according to claim 1, characterized in that the comparison of the new source code (6-2) with the relevant at least part of the output program component set (2-1) is a comparison of the new source code (6-2) with the output Model information (4-1) and the output structure description (8-1) is.

3. Method according to claim 2, characterized in that for comparison of the new source code (6-2) with the output model information (4-1) and the output structure description (8-1) from the new source code (6-2). 2) a first tree structure is constructed and this first tree structure on a the starting Model information (4-1) and the output structure description (8-1) corresponding second tree structure is mapped.

A method according to claim 2 or 3, characterized in that the step of generating the new structure description (8-2) and the new model information (4-2) comprises the steps of:

Generating the new structure description (8-2) on the basis of a first comparison of the at least one, first part (4-1, 8-1) of the components of the output program component set (2-1) with the new source code (6 -2), and

Generating the new model information (4-2) based on a second comparison of a second part (4-1) of the source program component set (2-1) with the new source code (6-2) and the new texture description (8-2) wherein the new model information (4-2) comprises structurally unrepresentable differences between the source program component set (2-1) and the new source code (6-2).

5. The method according to claim 1, wherein generating the new structure description (8-2) and the new model information (4-2) using the output model information (4-1), the output structure description (8 -1) and the new source code (6-2) is done automatically.

6. The method according to any one of claims 1 to 4, characterized in that generating the new structure description (8-2) and the new model information (4-2) using the output model information (4-1), the output structure description (8-1), the new source code (6-2) and a user interaction.

7. The method according to any one of the preceding claims, characterized in that after generating one of the new structure description (8-2) and the new model information (4-2) of the new source code (6-2) from the new model information (4 -2) is generated automatically according to the new structure description (8-2).