From: Mike F. <va...@ge...> - 2007-06-16 21:25:03
|
On Saturday 16 June 2007, Randy Dunlap wrote: > Mike Frysinger wrote: > > On Friday 15 June 2007, Randy Dunlap wrote: > >> On Wed, 13 Jun 2007 11:56:42 -0300 Ricardo Salveti de Araujo wrote: > >>> This patch fix the problem reported at bug "[ 1732287 ] mtest06/mmap1= =2Ec > >>> missing run_once check in do loop". > >>> > >>> What happens here is that before release ltp-full-20070531, the test > >>> had a default value for how much time it'd run, but when you passed > >>> nothing to it it'd run for just one iteration, without working with t= he > >>> default value (that's 24 hours). > >>> > >>> At the latest release, this was fixed but ltp runtest files is calling > >>> this test as before, passing nothing to it, and the test start running > >>> for 24 hours (instead of just one time). > >>> > >>> I had two options to fix this, one is ignoring the default run time a= nd > >>> the other is fixing how we call the test at runtest files. I did the > >>> second one because sound more correct to me than the first one (the > >>> test is now running for 3 minutes at runtest/* files). > >>> > >>> It has also a fix at the test output, also pointed by the bug. > >> > >> Thanks for the patches. This is worthy of a new ltp release IMO. > > > > i dont see why it should change the way ltp releases have always been > > done ... one snapshot a month > > Did mmap1 previously run for 24 hours as the default? you think this is the first test that broke in a release and had similar=20 misbehavior ? =2Dmike |