On Thu, Feb 06, 2020 at 05:40:42PM -0500, Peter Xu wrote: > On Wed, Feb 05, 2020 at 12:39:39PM -0500, Peter Xu wrote: > > On Wed, Feb 05, 2020 at 06:11:09PM +0100, Andrew Jones wrote: > > > On Wed, Feb 05, 2020 at 10:46:17AM -0500, Peter Xu wrote: > > > > On Wed, Feb 05, 2020 at 10:28:52AM +0100, Andrew Jones wrote: > > > > > On Tue, Feb 04, 2020 at 09:58:38PM -0500, Peter Xu wrote: > > > > > > Remove the clear_dirty_log test, instead merge it into the existing > > > > > > dirty_log_test. It should be cleaner to use this single binary to do > > > > > > both tests, also it's a preparation for the upcoming dirty ring test. > > > > > > > > > > > > The default test will still be the dirty_log test. To run the clear > > > > > > dirty log test, we need to specify "-M clear-log". > > > > > > > > > > How about keeping most of these changes, which nicely clean up the > > > > > #ifdefs, but also keeping the separate test, which I think is the > > > > > preferred way to organize and execute selftests. We can use argv[0] > > > > > to determine which path to take rather than a command line parameter. > > > > > > > > Hi, Drew, > > > > > > > > How about we just create a few selftest links that points to the same > > > > test binary in Makefile? I'm fine with compiling it for mulitple > > > > binaries too, just in case the makefile trick could be even easier. > > > > > > I think I prefer the binaries. That way they can be selectively moved > > > and run elsewhere, i.e. each test is a standalone test. > > > > Sure, will do. > > Or... Shall we still keep one binary, but by default run all the > supported logging mode in sequence in a single dirty_log_test? Say, > run "./dirty_log_test" will run all supported tests one by one; run > "./dirty_log_test -M LOG_MODE" will only run specific mode. > > With this patch it's fairly easy to achieve this too. > Works for me. Thanks, drew
