in Tue, May 18, 2010 at 7:04 PM, Jeff Johnson <n3...@mac.com> wrote: > > Looks pretty good. >
Thanks. It is possible to do better, of course. > > Couple of specific comments: > > 1) I wouldn't bother with AutoFu for splint. > > 2) Check for the ISPRAS tests in /usr/lib/rpm/bin too. Its distributed w > @rpm5.org. > I will (re)take a look at it for popt. > > Note that test2/test3 have never ever been usefully or reliably implemented > or maintained. > Which means that there's only test1 that has any meaning. > > The point is that I'm not sure that all the AutoFu baggage for > a test harness is largely a waste of time (imho) because its > just one KISSy script that invokes test1 53 different times. Sure. It is clear to me that improved coverage tests would be necessary > > Note that I _REALLY_ wish to avoid a test harness from HELL, > nothing more. If it floats your boat to use AutoFu for running > test.sh, that's fine. But if 50+ different invocations of > a single executable leads to 50+ different scripts/inputs/outputs > well, that's my de faction definition of HELL. Ok, Autofu was just my goal at this moment. It was what I promised. I'm sure that it is possible to find something more engineered for popt 2.0.The problem, as always, is the time, which, like life, is short. Best Regards ______________________________________________________________________ RPM Package Manager http://rpm5.org Developer Communication List rpm-devel@rpm5.org
