| 摘要 |
A test system controller receives a test code having N versions, selects n intermediate versions of the test code, and tests the test code using history revision testing. The controller runs the intermediate versions in parallel, and identifies a first passed version and a first failed version from the intermediate versions. The controller then selects a subset of the N versions, the subset including N′ versions between and including the first passed version and first failed version. The controller calculates a sample size, and determines whether this sample size is greater than n. Responsive to this, new intermediate versions are tested. The controller automatically performs at least a plurality of iterations of selecting and testing. Responsive to the sample size reaching a specified value, a plurality of N″ versions are run in parallel, such that a passed version and a failed version are identified. |
| 主权项 |
1. A method of history revision testing for locating a software property, the method comprising:
receiving a test code, the test code having N versions in chronological order, wherein N is a quantity of versions; selecting n intermediate versions of the test code, wherein n is a quantity of intermediate versions; and testing the test code using history revision testing that comprises:
running the intermediate versions in parallel;identifying a first passed version, the first passed version being a last intermediate version of the intermediate versions to pass the test;identifying a first failed version, the first failed version being a first intermediate version of the intermediate versions to fail the test, such that the first passed version and the first failed version are consecutive versions of the intermediate versions;selecting a subset of the N versions, the subset including N′ versions, the N′ versions comprising all versions of the N versions between the first passed version and the first failed version, including the first passed version and the first failed version;calculating a sample size, the sample size being calculated as a total number of N′ versions;determining whether the sample size is greater than n;responsive to determining that sample size is greater than n:
selecting n new intermediate versions of the test code from the N′ versions; andtesting the new intermediate versions using history revision testing that comprises:
running the new intermediate versions in parallel;identifying a second passed version, the second passed version being a last new intermediate version of the new intermediate versions to pass the test;identifying a second failed version, the second failed version being a first new intermediate version of the new intermediate versions to fail the test, such that the second passed version and the second failed version are consecutive versions of the new intermediate versions;selecting a subset of the N′ versions, the subset including N″ versions, the N″ versions comprising all versions of the N′ versions between the second passed version and the second failed version, including the second passed version and the second failed version;recalculating the sample size, the sample size being recalculated as a total number of N″ versions; andafter recalculating the sample size, determining whether the sample size is greater than n;automatically performing at least a plurality of iterations of selecting and testing, including selecting and testing the intermediate versions and the new intermediate versions; andresponsive to determining that the sample size is not greater than n:
running the N″ versions in parallel;identifying a third passed version, the third passed version being a last version of the N″ versions to pass the test; andidentifying a third failed version, the third failed version being a first version of the N″ versions to fail the test, such that the third passed version and the third failed version are consecutive versions of the N″ versions,wherein the third failed version is a first version of the N versions that includes the software property. |
