ERROR ASSIGNMENT FOR COMPUTER PROGRAMS

• US 20180365091A1
• Filed: 06/15/2017
• Published: 12/20/2018
• Est. Priority Date: 06/15/2017
• Status: Active Grant

• Current Assignee: Salesforce.com, Inc.
• Original Assignee: Salesforce.com, Inc.
• Inventors: Donaldson, J. Justin, Tarevern, Hormoz, Hameed, Sadiya, Srivastava, Siddharth, Jiang, Feifei
• Granted Patent: US 10,409,667 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 11/0706 the processing taking place on a specific hardware platform or in a specific software environment
• CPC Class Codes: G06F 11/0751 Error or fault detection not based on redundancy power supply failures G06F1/30; network fault management H04L41/06
• CPC Class Codes: G06F 11/0772 Means for error signaling, e.g. using interrupts, exception flags, dedicated error registers
• CPC Class Codes: G06F 11/0784 Routing of error reports, e.g. with a specific transmission path or data flow
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36
• CPC Class Codes: G06N 3/02 Neural networks
• CPC Class Codes: G06N 3/04 Architecture, e.g. interconnection topology
• CPC Class Codes: G06N 3/08 Learning methods

Abstract

An online system identifies an assignment for a computer program error indicated in an error message by applying an assignment model to token sequences identified in the error message. The error message includes a sequence of execution paths of the computer program. Each execution path indicates a function call active in computer memory when the error was generated. In other words, the error message allows tracking of the sequence of nested paths up to the point where the error was generated. In one example, the error message is a stack trace message that reports active stack frames in computer memory during the execution of the program.

Claims

1. A computer implemented method, comprising:

receiving an error message indicating error that occurred during execution of a computer program, wherein the error message includes a sequence of execution paths of the computer program;

extracting a plurality of token sequences in the error message, wherein each token sequence includes a sequence of token identifiers identified from the execution paths of the error message;

determining an assignment for the error by applying a machine-learned model to the extracted plurality of token sequences; and

providing the error message and corresponding error assignment record to a user associated with the assignment.

2. The computer implemented method of claim 1, wherein each execution path from the sequence of execution paths represents a function call active in computer memory when the error was generated.

3. The computer implemented method of claim 1, wherein determining the assignment for the error comprises:

determining, for each token sequence, an assignment for the token sequence by applying a machine-learned model to the token sequence; and

selecting an assignment having the highest likelihood given the corresponding plurality of token sequences.

4. The computer implemented method of claim 3, wherein the machine-learned model is a neural network model, and determining the assignment for each token sequence comprises:

generating an output by applying the same transformation of the neural network model to each token identifier in the token sequence; and

determining the assignment for the token sequence based on the generated output.

5. The computer implemented method of claim 1, further comprising encoding each token identifier from the plurality of token sequences into a one-hot encoded vector.

6. The computer implemented method of claim 1, wherein each token sequence in the plurality of token sequences includes token identifiers corresponding to a same programming element type.

7. The computer implemented method of claim 6, wherein each token sequence includes token identifiers corresponding to an object-oriented class programming element.

8. The computer implemented method of claim 1, wherein the error message is a stack trace message that reports active stack frames in computer memory during execution of the computer program.

9. A non-transitory computer readable storage medium comprising computer executable code that the when executed by one or more processors causes the one or more processors to perform operations comprising:

receiving an error message indicating error that occurred during execution of a computer program, wherein the error message includes a sequence of execution paths of the computer program;

identifying a plurality of token sequences in the error message, wherein each token sequence includes a sequence of token identifiers identified from the execution paths of the error message;

determining an assignment for the error by applying a machine-learned model to the identified plurality of token sequences; and

providing the error message to a user associated with the assignment.

10. The computer readable storage medium of claim 9, wherein each execution path from the sequence of execution paths represents a function call active in computer memory when the error was generated.

11. The computer readable storage medium of claim 9, wherein determining the assignment for the error comprises:

determining, for each token sequence, an assignment for the token sequence by applying a machine-learned model to the token sequence; and

selecting an assignment having the highest likelihood among the assignments determined for the plurality of token sequences.

12. The computer readable storage medium of claim 11, wherein the machine-learned model is a neural network model, and determining the assignment for each token sequence comprises:

generating an output by applying the same transformation of the neural network model to each token identifier in the token sequence; and

determining the assignment for the token sequence based on the generated output.

13. The computer readable storage medium of claim 9, further comprising encoding each token identifier in the plurality of token sequences into a one-hot encoded vector.

14. The computer readable storage medium of claim 9, wherein each token sequence in the plurality of token sequences includes token identifiers corresponding to a same programming element type.

15. The computer readable storage medium of claim 14, wherein each token sequence includes token identifiers corresponding to an object-oriented class programming element.

16. The computer readable storage medium of claim 9, wherein the error message is a stack trace message that reports active stack frames in computer memory during execution of the computer program.

17. A system comprising:

one or more computer processors; and

a non-transitory computer readable storage medium comprising computer executable code that the when executed by the one or more processors causes the one or more processors to perform operations comprising;

receiving an error message indicating error that occurred during execution of a computer program, wherein the error message includes a sequence of execution paths of the computer program;

identifying a plurality of token sequences in the error message, wherein each token sequence includes a sequence of token identifiers identified from the execution paths of the error message;

determining an assignment for the error by applying a machine-learned model to the identified plurality of token sequences; and

providing the error message to a user associated with the assignment.

18. The system of claim 17, wherein each execution path from the sequence of execution paths represents a function call active in computer memory when the error was generated.

19. The system of claim 17, wherein determining the assignment for the error comprises:

determining, for each token sequence, an assignment for the token sequence by applying a machine-learned model to the token sequence; and

selecting an assignment having the highest number among the assignments determined for the plurality of token sequences.

20. The system of claim 19, wherein the machine-learned model is a neural network model, and determining the assignment for each token sequence comprises:

generating an output by applying the same transformation of the neural network model to each token identifier in the token sequence; and

determining the assignment for the token sequence based on the generated output.