On Thu, Jul 5, 2018 at 6:07 PM Satish Balay <ba...@mcs.anl.gov> wrote:
> Ok - there is some confusion here. > > The test harness does not use targets in > examples/[tests,tutorials]/makefile. > > However we want to keep the functionality of us/users compiling > individual examples manually - without going through the test harness. > > i.e the following should continue to work: > > cd src/ksp/ksp/examples/tutorials > make ex2 > I believe that only relies on a) Having compile targets, or a generic example compile target b) the compile rules in the toplevel makefiles I do not think Jed objects to either. Matt > On Thu, 5 Jul 2018, Jed Brown wrote: > > > Satish Balay <ba...@mcs.anl.gov> writes: > > > > > On Thu, 5 Jul 2018, Jed Brown wrote: > > > > > >> When can we delete the legacy test system? Are we currently using it > > >> anywhere? > > > > > > [Don't know exactly which parts we would delete] - but the new targets > > > cover ex*[f,f90] type examples - and not anything else. > > > > Which targets are you referring to? I don't see any build targets in > > the source directories. > > ex%f90: etc targets in lib/petsc/conf/test > > > There are a few run targets that I don't > > understand. We could delete lib/petsc/conf/test and any top-level > > legacy targets. > > well lib/petsc/conf/test enables the above feature of 'make ex2' > > > > > > So all other examples that don't fit this string format still use the > > > old targets. > > > > Where are those? They evidently aren't part of the nightly tests, thus > > need to be updated. > > [With the above context] - I was refering to - say > spectraladjointassimilation, burgers_spectral etc targes in > src/tao/unconstrained/examples/tutorials/makefile > > Sure - this is not part of the test harness and nightlybilds. > > > > > Also they are still useful to users [to create their own makefiles] - > > > as we don't have an equivalent simiple replacement yet.. > > > > I don't mind makefiles sitting in source directories. They're basically > > just includes and a list of source files right now, and we can remove > > the lists of source files. > > They also list dependencies compile dependencies [for the above purpose] > > Satish > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener https://www.cse.buffalo.edu/~knepley/ <http://www.caam.rice.edu/~mk51/>
