Multidimensional spatial searching for identifying duplicate crash dumps

1. A method of identifying duplicate crash dumps in a computer system, the method comprising:

receiving a first crash dump caused by an application crash;

extracting, from the first crash dump, a first function signature of a function that caused the first crash dump;

searching a datastore of crash dumps for function signatures that substantially match the first function signature, wherein the searching comprises:

performing an approximate string-match between each of the function signatures and the first function signature;

comparing the approximate string-matches to a first threshold;

performing an exact string match between each of the function signatures and the first function signature;

comparing the exact string-matches to a second threshold;

combining weighted results of the approximate string-match with weighted results of the exact string match to generate match scores for each of the function signatures;

comparing the match scores to a third threshold; and

identifying the function signatures that substantially match the first function signature based on:

the comparing of the approximate string-matches to the first threshold;

the comparing of the exact string-matches to the second threshold; and

the comparing of the match scores to the third threshold.

2. The method of claim 1 , further comprising:

extracting call stacks from the crash dumps in the datastore of crash dumps;

extracting the function signatures from the call stacks.

3. The method of claim 1 , further comprising:

returning function signatures with string matching scores above a threshold score.

4. The method of claim 1 , wherein the datastore of crash dumps comprises a database that stores crash dump information from previous application crashes.

5. The method of claim 1 , wherein extracting the first function signature from the crash dump comprises selecting a top-most function signature in a call stack of the first crash dump.

6. The method of claim 1 , wherein extracting the first function signature from the crash dump comprises identifying unique function signatures that are called when an application crashes.

7. The method of claim 1 , further comprising, after receiving the first crash dump, sanitizing a call stack in the first crash dump by removing all non-application function calls.

8. The method of claim 1 , wherein the exact string match comprises a Longest Common Substring search.

9. The method of claim 1 , wherein the approximate string match comprises calculating a Levenshtein score.

10. The method of claim 1 , further comprising:

causing results to be displayed on the display device, wherein the results comprise a table including each of the function signatures and the associated matching scores.

11. A system for identifying duplicate crash dumps, the system comprising:

a crash dump data store, wherein:

the crash dump data store includes a plurality of crash dumps; and

the plurality of crash dumps comprises function signatures;

a computer system running an application, wherein when the application crashes, a first crash dump is triggered; and

a function matching module operating on a server that is in communication with the computer system, wherein:

the function matching module receives the first crash dump from the computer system;

the function matching module extracts, from the first crash dump, a first function signature of a function that caused the first crash dump;

the function matching module searches the plurality of crash dumps from the crash dump data store for function signatures that substantially match the first function signature by:

performing an approximate string-match between each of the function signatures and the first function signature;

comparing the approximate string-matches to a first threshold;

performing an exact string match between each of the function signatures and the first function signature;

comparing the exact string-matches to a second threshold;

combining weighted results of the approximate string-match with weighted results of the exact string match to generate match scores for each of the function signatures;

comparing the match scores to a third threshold; and

identifying the function signatures that substantially match the first function signature based on:

the comparing of the approximate string-matches to the first threshold;

the comparing of the exact string-matches to the second threshold; and

the comparing of the match scores to the third threshold.

12. The system of claim 11 , wherein the function matching module further:

extracts call stacks from the crash dumps in the datastore of crash dumps;

extracts the function signatures from the call stacks.

13. The system of claim 11 , wherein the function matching module further:

returns function signatures with string matching scores above a threshold score.

14. The system of claim 11 , wherein the datastore of crash dumps comprises a database that stores crash dump information from previous application crashes.

15. The system of claim 11 , wherein extracting the first function signature from the crash dump comprises selecting a top-most function signature in a call stack of the first crash dump.

16. The system of claim 11 , wherein extracting the first function signature from the crash dump comprises identifying unique function signatures that are called when an application crashes.

17. A non-transitory, computer-readable medium comprising instructions which, when executed by one or more processors, causes the one or more processors to perform operations comprising:

receiving a first crash dump caused by an application crash;

extracting, from the first crash dump, a first function signature of a function that caused the first crash dump;

searching a datastore of crash dumps for function signatures that substantially match the first function signature, wherein the searching comprises:

performing an approximate string-match between each of the function signatures and the first function signature;

comparing the approximate string-matches to a first threshold;

performing an exact string match between each of the function signatures and the first function signature;

comparing the exact string-matches to a second threshold;

combining weighted results of the approximate string-match with weighted results of the exact string match to generate match scores for each of the function signatures;

comparing the match scores to a third threshold; and

identifying the function signatures that substantially match the first function signature based on:

the comparing of the approximate string-matches to the first threshold;

the comparing of the exact string-matches to the second threshold; and

the comparing of the match scores to the third threshold.

18. The non-transitory computer-readable memory of claim 17 , wherein the exact string match comprises a Longest Common Substring search.

19. T's he non-transitory computer-readable memory of claim 17 , wherein the approximate string match comprises calculating a Levenshtein score.

20. The non-transitory computer-readable memory of claim 17 , wherein the instructions cause the one or more processors to perform further operations comprising:

causing results to be displayed on the display device, wherein the results comprise a table including each of the function signatures and the associated matching scores.