John, There were instances where the fsck would catch an error in the file system, correct it and throw and error. This would get logged in the CLIENT.log file as an error but the overall test would fail. http://b/issue?id=2593141&cookieId=2010118161057631
Adding in this code seems to force the test to fail as well. --- Cheers! Akshay Lal On Wed, Apr 28, 2010 at 4:09 PM, John Admanski <[email protected]> wrote: > Wed, Apr 28, 2010 at 3:25 PM, John Admanski <[email protected]> wrote: > >> On Wed, Apr 28, 2010 at 3:20 PM, Martin Bligh <[email protected]> wrote: >> >>> On Wed, Apr 28, 2010 at 3:18 PM, Curt Wohlgemuth <[email protected]> >>> wrote: >>> > It would be nice if we didn't have to go add the fsck call to each >>> > test. Since the partitions module is used for all FS tests, having it >>> > there for use by all makes sense to me. >>> >>> Right, this is what the groups code is meant to handle, it makes >>> a group from: >>> >>> mkfs >>> mount >>> run_test >>> umount >>> fsck >>> >>> and if any of those fails it's mean to represent the test "group" as >>> a failure. >>> >> >> Well, non Test* exceptions generally aren't caught by either run_test or >> run_group; the implication is that something worse happened and so only the >> top-level job handler catches it (and aborts the job). >> >> I don't believe there's any special class of exception which is caught by >> the run_group handler (and not re-raised). >> >> -- John >> > > Overall I don't see any reason to not apply this, it seems to comply with > the way test exceptions and run_test and run_group work right now. That > said, I also don't see how this makes any difference to the way things work; > I don't see where in the call stack the TestError is going to make things > work differently than the original raise. > > -- John >
_______________________________________________ Autotest mailing list [email protected] http://test.kernel.org/cgi-bin/mailman/listinfo/autotest
