From: Robin G. <rg...@bl...> - 2005-01-14 12:50:57
|
We are using LTP to validate a uClinux port that we are doing to a new processor. We have seen issues that if the LTP is run through (in the "correct" order) - things work fine, and all the tests pass. Then, as a stress test, we continue to run LTP out of order (random tests that passed previously), and now we see failures on tests that have previously passed. Unfortunately for us, this is pretty repeatable. Has anyone else seen an test order dependency in LTP? (or do we have some lurking issue in our port somewhere)... Thanks -Robin P.S. I am wondering how this would link into the reporting/query techniques that are being discussed? |