On 10/05/06, Tim Dwyer <Tim.Dwyer@...38...> wrote:
Interesting! This particular test code (written by Peter Moulder) was set to fail if it didn't complete in 30 seconds (hopefully catching potential infinite loops). It's possible that if you're running it on a really slow machine
... I think that that was indeed the case. It is not so much that the machine was slow, but that it was doing other things as well. My suggestion would be to assign a time out for Count of n based on the time taken for Count of n-1.
Ben Fowler wrote:
I get this message:
Testing removeRectangleOverlap(zero gaps)... eg0...OK eg2...OK eg3...OK eg4...OK eg5...OK olap6...OK loop0...OK loop1...OK loop2...OK random ints with 0 rectangles...OK random ints with 1 rectangles...OK random ints with 2 rectangles...OK random ints with 3 rectangles...OK random ints with 4 rectangles...OK random ints with 5 rectangles...OK random ints with 6 rectangles...FAIL: removeoverlap/remove_rectangle_overlap-test
Is it worth sending such a report to this list. 'make check' did not report any problems lst time I ran it.
The tests with the 'random ints' actually take a bit longer than any of the others and sometimes I get a message about an Alarm Clock.
Does this indicate something fixable at my end, or a possible problem with the testing system... Or anything in particular?