Stack trace generated code compared with database to find error resolution information

[0001] This invention relates generally to identifying a program error in a computer program, and more particularly to automatically identifying a program error in a computer program.

[0002] A computer program may generate a program error during execution. It may be desirable to identify the program error and to determine a resolution for it. In some cases, the program error may have been previously identified and resolved. For example, the same program or a different program, such as a different version of the same program, previously executed within a computer network may have previously generated the same program error. In response to a new program error, a program stack trace may be automatically generated, representing a path to the source of the new program error in the program that generated the new program error. To reduce potentially duplicative work in resolving a new program error, it may be desirable to determine whether the same program error has been generated previously and, if so, whether a resolution for the program error is available. To identify the new program error, a user such as a program developer may be required to manually search previously generated program stack traces to determine whether a matching program stack trace exists for a previous program error, indicating that the previous program error is the same as the new program error. This process is often tedious, costly, and prone to human error.

[0003] According to the present invention, problems and disadvantages associated with previous techniques for identifying and resolving a program error in a computer program may be reduced or eliminated.

[0004] In one embodiment, a method for automatically identifying a program error in a computer program includes receiving a program error message indicating that a new program error has occurred during execution of the program and generating a program stack trace for the new program error indicating the source of the new program error within the program. A first operation is applied to a portion of the program stack trace to generate a first numerical code associated with the new program error. The first numerical code associated with the new program error is compared to one or more stored first numerical codes associated with previously identified program errors to determine whether the new program error is the same as a previously identified program error. If so, stored information associated with resolving the previously identified program error is provided for use in resolving the new program error.

[0005] Certain embodiments of the present invention may provide one or more technical advantages. For example, certain embodiments may uniquely identify a program error using one or more numerical codes generated according to a program stack trace for the program error. Uniquely identifying program errors according to such numerical codes may allow a new program error to be readily and automatically compared with previously identified program errors to determine whether information that may be used to resolve the new program error already exists. This may allow a user such as a program developer to avoid duplicative work, which may reduce tedium, lower development costs, and improve accuracy. In certain embodiments, multiple numerical codes may be generated for each program error according to the associated program stack trace, which may allow program errors generated in different versions of a program to be identified as the same. Systems and methods incorporating one or more of these or other technical advantages may be well suited for identifying and resolving program errors. Certain embodiments may provide all, some, or none of the above advantages. Certain embodiments may provide one or more other technical advantages, one or more of which may be readily apparent to those skilled in the art from the figures, descriptions, and claims included herein.

[0012] FIG. 1 illustrates an example system for automatically identifying a program error in a computer program. System 10 includes one or more client systems 12 and one or more server systems 14 coupled using a network 16. Client system 12 and server system 14 each operate on one or more computers at one or more locations. Each computer may include one or more suitable input devices, output devices, processors and associated memory, mass storage media, communications interfaces, and any other components that may be appropriate according to operation of system 10. Network 16 may include one or more local area networks (LANs), metropolitan area networks (MANs), wide area networks (WANs), portions of the global computer network known as the Internet, or any other appropriate wireline, optical, wireless, or other links.

[0013] Components of server system 14 may include a web server 18, an application server 24, an application 26 (which may be one of multiple applications associated with application server 24), and a database 28. In one embodiment, application server 24 includes a WEBLOGIC application server, WEBSPHERE application server, or other suitable application server and application 26 includes a JAVA application. However, the present invention contemplates application 26 with or without application server 24. Using a web browser or otherwise, a client system 12 may submit a request to web server 18 of server system 14 for server system 14 to execute a computer program. Web server 18 may communicate the request to application server 24, prompting application server 24 to initiate execution of the program. During execution of the program, one or more program errors may occur. Application 26 may receive a program error message, indicating that a new program error has occurred during execution of the program. Application 26 may generate a program stack trace for the new program error, which may indicate the source of the new program error within the program. Application 26 may generate one or more numerical codes associated with the new program error, compare the numerical codes associated with the new program error to one or more stored numerical codes associated with previously identified program errors, and if a match is determined provide stored information associated with resolving the matching previously identified program error for use in resolving the new program error.

[0014] Although a particular client-server environment is described as an example, the present invention contemplates any suitable computing environment. Furthermore, although application 26 is described as performing certain operations, the present invention contemplates these operations being performed using any suitable component or components associated with server system 14 or another appropriate computer system. For example, certain operations may be performed by application 26 (e.g., by one of multiple processes of application 26), by another application (e.g., by one of multiple processes of the other application), by a developer or other user associated with system 10, or in any other suitable manner. Moreover, although computer programs are described, the present invention contemplates identifying and resolving program errors in any suitable software component or code segment, whether or not referred to as a “program” in the particular language in which the software component or code segment is written.

[0015] FIG. 2 illustrates an example program stack trace 40, which may indicate the source of a new program error within a program. Program stack trace 40 may be a standard JAVA program stack trace or any other suitable program stack trace according to particular needs. In one embodiment, program stack trace 40 may be described as a path to the computer code in the program that generated the program error. For example, program stack trace 40 may indicate one or more class methods or other suitable methods 42 in the program that generated the program error. For each such method 42, program stack trace 40 may include a line number 44 of the code in method 42 where the program error was generated.

[0016] Returning to FIG. 1, application 26 may generate one or more numerical codes associated with the new program error. In one embodiment, application 26 may compare the numerical codes associated with the new program error to corresponding stored numerical codes associated with previously identified program errors to determine whether the new program error is the same as at least one of the previously identified program errors. However, this operation or operations may be performed by application 26 (e.g., by one of multiple processes of application 26), by another application (e.g., by one of multiple processes of the other application), by a developer or other user associated with system 10, or in any other suitable manner depending on the embodiment. The stored numerical codes may be stored in database 28 or at any other suitable location accessible to application 26. The stored numerical codes may be stored by previous operations of application 26 (e.g., by one of multiple processes of application 26), by another application (e.g., by one of multiple processes of the other application), by a developer or other user associated with system 10, or in any other suitable manner.

[0017] If a numerical code associated with the new program error is the same as a stored numerical code associated with a previously identified program error, application 26 may provide stored information associated with resolving the matching previously identified program error for use in resolving the new program error. For example, the stored information may include error-free code, explanatory text, a link to such code or text, a uniform resource locator (URL) or other address of such code or text, or any other suitable stored information. The stored information may have been stored by previous operations of application 26 (e.g., by one of multiple processes of application 26), by another application (e.g., by one of multiple processes of the other application), by a developer or other user associated with system 10, or in any other suitable manner depending on the embodiment. In one embodiment, if no numerical code associated with the new program error is the same as any stored numerical code associated with a previously identified program error, application 26 may store the one or more numerical codes associated with the new program error in database 28 for use in identifying subsequent program errors.

[0018] Certain embodiments of the present invention may provide one or more technical advantages. For example, certain embodiments may uniquely identify a program error using one or more numerical codes generated according to program stack trace 40 for the program error. Uniquely identifying program errors according to such numerical codes may allow a new program error to be readily and automatically compared with previously identified program errors to determine whether information that may be used to resolve the new program error already exists. This may allow a user such as a program developer to avoid duplicative work, which may reduce tedium, lower development costs, and improve accuracy. In certain embodiments, multiple numerical codes may be generated for each program error according to associated program stack trace 40, which may allow program errors generated in different versions of a program to be identified as the same.

[0019] FIGS. 3A–3C illustrate example computer programs 50, which server system 14 may execute in response to a request from client system 12. Programs 50 may each include one or more functions 42, which in a particular example may include func1, func2, func3, func_jaa, and main. While specific names and formats are given to functions 42 in programs 50, those skilled in the art will appreciate that any suitable names and formats may be used. Additionally, while specific programs 50 are described, the present invention contemplates any suitable programs 50.

[0020] When a program 50 is executing, certain functions 42 within program 50 may generate program errors as they are executing, causing program 50 to generate a program error. As an example, func_jaa in program 50a may generate a program error because func_jaa attempts to access an object that does not exist (i.e. points to null). Thus, when function main in program 50a calls func1, which calls func2, which calls func3, which calls func_jaa, a program error may be generated. Function main may contain an expression catch (Exception E), which may receive a program error message indicating that a new program error has occurred in func_jaa. Function main may also contain a function call addLogEntry(stackCode(e)), which may generate program error output, possibly including program stack trace 40 and the one or more numerical codes associated with the new program error.

[0021] FIGS. 4A–4C illustrate example program error output 70 of computer programs 50, which may be generated if a program 50 generates a program error and may be used to automatically identify the program error. Program error outputs 70a–70c illustrated in FIGS. 4A–4C, respectively, correspond to programs 50a–50c illustrated in FIGS. 3A–3C, respectively. For example, program error output 70a is generated in response to a program error occurring during execution of program 50a. While specific program error output 70 is described, those skilled in the art will appreciate that any program error output suitable for automatically identifying a program error in a program 50 may be used.

[0022] Program error output 70 may include program stack trace 40 indicating the source of the program error within program 50. Program error output 70 may also include one or more numerical codes 74 associated with the program error. A numerical code 74 may provide a numerical representation of program stack trace 40 and may uniquely identify the program error associated with program stack trace 40. In one embodiment, a numerical code 74 for a program error may be calculated by applying an operation that generates a substantially unique code to program stack trace 40 for the program error. In one embodiment, for example, a numerical code 74 for a program error may be calculated by applying a String.hashCode( ) or an Object.hashCode( ) method to program stack trace 40. For example, the operation may map an infinite universe of inputs to an integer x such that −2,147,493,649<x<2,147,493,649. The operation may be applied to all or a portion of program stack trace 40. It may be desirable to generate multiple numerical codes 74 to ensure that a new program error will be found to match a previously identified program error even though programs 50 that generated the program errors may not be identical. For example, program error output 70 may include an exact match code 76, a trace code 78, a place code 80, or any other suitable numerical codes 74.

[0023] In one embodiment, exact match code 76 may be a numerical code 74 computed when an operation that generates a substantially unique code is applied to the entire program stack trace 40 for a program error. As an example, an exact match code 76 associated with a new program error may be the same as an exact match code 76 associated with a previously identified program error when program 50 that generated the new program error is substantially identical to program 50 that generated the previously identified program error. Trace code 78 may be a numerical code 74 computed when an operation that generates a substantially unique code is applied to program stack trace 40 for a program error after deleting all line numbers 44 from program stack trace 40. As an example, a trace code 78 associated with a new program error may be the same as a trace code 78 associated with a previously identified program error when program 50 that generated the new program error has more or fewer lines of code than, but is otherwise identical to, program 50 that generated the previously identified program error. Place code 80 may be a numerical code 74 computed when an operation that generates a substantially unique code is applied to only specified lines, such as the first three lines for example, of program stack trace 40. As an example, a place code 80 associated with a new program error may be the same as a place code 80 associated with a previously identified program error when program 50 that generated the new program error includes one or more differently named functions 42 than program 50 that generated the previously identified program error.

[0024] As just an example, assume program 50a is executed and generates a program error message that is received by application 26. Application 26 may generate program error output 70a, possibly including program stack trace 40a indicating that func_jaa in program 50a is the source of the program error. Application 26 may then generate one or more numerical codes 74a providing unique identifiers for program stack trace 40a and thus for the new program error. Numerical codes 74a may include exact match code 76a, trace code 78a, place code 80a, or any other suitable numerical codes according to particular needs. Assume in this example that numerical codes 74a are stored in database 28. For example, the program error may be a new program error, such that none of the one or more numerical codes 74 for the program error match any stored numerical codes 74 of any previously identified program errors.

[0025] Continuing the example, assume program 50b is executed some time after numerical codes 74a are stored in database 28. The same code in func_jaa of program 50a that caused program 50a to generate a program error message is included in func_jaa of program 50b. Thus, when program 50b is executed, program 50b also generates a program error message that is received by application 26. Application 26 may generate program error output 70b, possibly including program stack trace 40b indicating that func_jaa in program 50b is the source of the program error. Application 26 may then generate one or more numerical codes 74b providing unique identifiers for program stack trace 40b and thus for the new program error. Thus, the same code in the same method 42 (func_jaa) that generated the program error in program 50b also generated the program error in program 50a. However, because in program 50b additional lines of code have been added to func_jaa, one or more line numbers 44b in program error output 70b may differ from corresponding line numbers 44a in program error output 70a. Thus, exact match code 76b may differ from exact match code 76a. However, trace code 78b may match trace code 78a because trace code 78 is computed after deleting line numbers 44 from program stack trace 40. Application 26 may then determine that trace code 78b associated with the new program error of program 50b is the same as stored trace code 78a associated with the previously identified program error of program 50a. Accordingly, application 26 may access database 28 to provide stored information associated with resolving the previously identified program error of program 50a for use in resolving the new program error of program 50b.

[0026] Continuing with the example, assume program 50c is executed some time after numerical codes 74a are stored in database 28. The same code in func_jaa of program 50a that caused program 50a to generate a program error message is included in program 50c. Thus, when program 50c is executed, program 50c generates a program error message that is received by application 26; however, the method 42 in program 50c containing the program error-generating line of code is named func_kbb instead of func_jaa. Application 26 may generate program error output 70c, possibly including program stack trace 40c indicating that func_kbb in computer program 50c is the source of the program error. Application 26 may then generate one or more numerical codes 74c providing unique identifiers for program stack trace 40c and thus for the new program error. Because in program 50c the name of the method 42 containing the error-generating line of code is different than the name of the method 42 in program 50a (func_kbb instead of func_jaa), one or more functions 42c in program stack trace 40c may differ from corresponding functions 42a in program stack trace 40a. Thus, exact match code 76c and trace code 78c may differ from exact match code 76a and trace code 78c, respectively. However, place code 80c may match place code 80a because place code 80 is computed using only a specified number of lines, such as the first three lines of program stack trace 40, which contain the differing function name. Application 26 may then determine that place code 80c associated with the new program error of program 50c is the same as stored place code 80a associated with the previously identified program error of program 50a. Accordingly, application 26 may access database 28 to provide stored information associated with resolving the previously identified program error of program 50a for use in resolving the new program error of program 50c.

[0027] FIG. 5 illustrates an example method for automatically identifying a program error in a computer program 50. At step 100, application 26 of server system 14 receives a program error message indicating that a new program error has occurred during execution of program 50. At step 102, application 26 generates program stack trace 40 for the new program error, program stack trace 40 indicating the source of the new program error within program 50. At step 104, application 26 generates one or more numerical codes 74 associated with the new program error. For example, application 26 may generate exact match code 76, trace code 78, place code 80, or any other suitable numerical codes 74. At step 106, application 26 compares a first numerical code 74 associated with the new program error, such as exact match code 76 for example, to one or more corresponding stored first numerical codes 74 associated with previously identified program errors to determine whether the new program error is the same as at least one of the previously identified program errors. The stored numerical codes may be stored by previous operations of application 26 (e.g., by one of multiple processes of application 26), by another application (e.g., by one of multiple processes of the other application), by a developer or other user associated with system 10, or in any other suitable manner depending on the embodiment.

[0028] At step 108, if the first numerical code 74 associated with the new program error is the same as the first numerical code 74 of a previously identified program error, application 26 may provide stored information associated with resolving the matching previously identified program error for use in resolving the new program error at step 110. For example, the stored information may include error-free code, explanatory text, a link to such code or text, a uniform resource locator (URL) or other address of such code or text, or any other suitable stored information. The stored information may have been stored by previous operations of application 26 (e.g., by one of multiple processes of application 26), by another application (e.g., by one of multiple processes of the other application), by a developer or other user associated with system 10, or in any other suitable manner depending on the embodiment.

[0029] If the first numerical code 74 associated with the new program error is not the same as the first numerical code 74 of a previously identified program error at step 108, application 26 determines whether there is a next numerical code 74 to compare to corresponding stored numerical codes 74 of previously identified program errors at step 112. If a next numerical code 74 has been generated at step 112, such as trace code 76 for example, the method returns to step 106, where application 26 compares that numerical code 74 associated with the new program error to corresponding stored numerical codes 74 associated with previously identified program errors. The method may continue in this manner until all numerical codes 74 for the new program error generated at step 104 have been compared to corresponding stored numerical codes 74 generated for previously identified program errors. For example, in one embodiment, exact match code 76, trace code 78, and place code 80 are compared in that order. If a next numerical code 74 has not been generated at step 112, the new program error is an entirely new program error at step 114 and application 26 may store the one or more numerical codes 74 associated with the new program error at step 116 for use in identifying subsequent program errors.

[0030] The present invention contemplates the steps of the method being performed in any suitable order. For example, the first numerical code 74 associated with the new program error may be generated and compared to first numerical codes 74 associated with previously identified program errors before any other numerical codes 74 associated with the new program error are generated, a next numerical code 74 being generated only if no match is determined. As another example, all numerical codes 74 associated with the new program error (e.g., exact match code 76, trace code 78, and place code 80) may be compared to all corresponding stored numerical codes 74 associated with previously identified program errors at substantially the same time. As another example, even if a numerical code 74 associated with the new program error is the same as a corresponding stored numerical code 74 associated with a previously identified program error, one or more additional numerical codes 74 associated with the new program error may be compared to corresponding stored numerical codes 74 associated with previously identified program errors to verify or otherwise confirm that the new program error is the same as the previously identified program error.

[0031] Although the present invention has been described with several embodiments, diverse changes, substitutions, variations, alterations, and modifications may be suggested to one skilled in the art, and it is intended that the invention encompass all such changes, substitutions, variations, alterations, and modifications as fall within the spirit and scope of the appended claims.