Merge branch 'scott/test-harness-timing2'

Report wall clock time of total tests

The timing report could be a bit confusing especially when running in
parallel. This separates out a wall clock time and timing of individual
tests. At presen

Report wall clock time of total tests

The timing report could be a bit confusing especially when running in
parallel. This separates out a wall clock time and timing of individual
tests. At present, using the 'real' time instead of 'sys+user' so this
still may not be optimal.

show more ...

Merge branch 'jed/gmakefile-paths' [PR #885]

* jed/gmakefile-paths:
makefile: add + to recursive recipes so that "make test" does not warn about -j
gmakefile: use absolute path for SL_LINKER_FUN

Merge branch 'jed/gmakefile-paths' [PR #885]

* jed/gmakefile-paths:
makefile: add + to recursive recipes so that "make test" does not warn about -j
gmakefile: use absolute path for SL_LINKER_FUNCTION
gmakefile: avoid replication of shared library recipe
makefile: use + in common recursive make invocations
gmakegentest: support localrunfiles naming individual files in subdirectories
gmake: remove src/ from build (PETSC_ARCH/obj) and test paths
gmake tests: enable testing prefix install without write access

show more ...

Merge remote-tracking branch 'origin/knepley/fix-snes-ex69' into tisaac/feature-dmfield

* origin/knepley/fix-snes-ex69: (311 commits)
PetscFECreateDefault: fix custom Fortran interface to use MPI_

Merge remote-tracking branch 'origin/knepley/fix-snes-ex69' into tisaac/feature-dmfield

* origin/knepley/fix-snes-ex69: (311 commits)
PetscFECreateDefault: fix custom Fortran interface to use MPI_Comm
SNES ex69: Updated test output
SNES ex69: Removed all mantle stuff and change fields to constants - Also use dsymutil for executables
SNES ex69: Removed special partition
SNES ex69: Fixed null space creation - Null space locks its vectors
PetscFE: PetscFECreateDefault() now takes MPI_Comm instead of DM
cleanup very ugly PCGAMG manual page
minor clarifications for VecScatterCreate() manual page
gcov: PETSC_ARCH is required onlyfor -run_gcov - and not for -merge_gcov [where it doesn't exist]
make clearer in help statements what --with-external-packages-dir is for.
Website: Added news entry for PETSc User Meeting 2018 on main page.
CUDA: Further test output fixes for mat_tests-ex2
Turn off elemental test for parmetis running since this does not work on batch systems.
3.8 Changes: KSPCHEBYSHEV
CUDA: Fixed most test errors due to MatMultTranspose*
website: update urls
website: fix Jason's name to match pic. Also add Todd's pic
Dev manual: copyedit
DMPlexDistributeLabels: Propagate the output flag
cuda/cusp: update minimum supported versions of cuda to 7.5 and cusp to 0.5.0
...

show more ...

gmake: remove src/ from build (PETSC_ARCH/obj) and test paths

This yields shorter targets to type, since formerly src/ was the only
directory in these paths.

gmake tests: enable testing prefix install without write access

Standard configurations should work as before, except that test
executables are once again named using relative paths (as documented i

gmake tests: enable testing prefix install without write access

Standard configurations should work as before, except that test
executables are once again named using relative paths (as documented in
"make help-test"). For prefix installs, testing is performed as

export PETSC_DIR=/path/to/prefix PETSC_ARCH=
make TESTDIR=/tmp/petsc-tests -f $PETSC_DIR/share/petsc/examples/gmakefile.test test

where TESTDIR is some location that is writable.

A side-effect of this change is that the generated
$PETSC_ARCH/lib/petsc/conf/testfiles has moved to $TESTDIR/testfiles.

show more ...

Merge remote-tracking branch 'origin/jed/fix-fortran-circular-deps'

Merge remote-tracking branch 'origin/scott/test-harness-timing'

tests: print the list of 5 longest running tests [within 'make alltests']

Cleanup Fortran circular dependencies produced by gfortran

When a module is defined and used in the same compilation unit, the
gfortran dependencies list the module file as both a target and a
prere

Cleanup Fortran circular dependencies produced by gfortran

When a module is defined and used in the same compilation unit, the
gfortran dependencies list the module file as both a target and a
prerequisite. Here we chop down the dependency statements to only hold
the %.o target that is referenced by our makefiles.

show more ...

Merge branch 'scott/test-harness-installfixes'

Merge branch 'scott/test-harness-minorfixes'

Minor fixes

Make sure all testflags are documented.
Add -f to an rm -r to remove error warnings.

Fix for report_tests from install dirs

Improvements for the install

The test harness has to operate in 3 modes:
1. Normal development mode
2. Test install from source dir (make allgtests)
3. Test install from install dir
See config

Improvements for the install

The test harness has to operate in 3 modes:
1. Normal development mode
2. Test install from source dir (make allgtests)
3. Test install from install dir
See config/gmakegentest.py documentation for how these differs.
This sorts out the paths for all 3 cases.

A big change is making gmakefile.test use absolute paths
rather than the relative paths used originally and in the normal
gmakefile. This simplifies a lot of the development, including the
previous method of doing a lot of modifications for the installs.

`make test` is still not hooked up to the new test system so testing was
done with allgtest.

show more ...

Merge branch 'scott/ksp-ksp-tutorials'

tests: some example includes can be in the same location of the sourcefile. so specify -Isrc in fortran compile targets

This fixes fortran example compiles on windows.

modify gfortran mmd fix (.d files)

Nightlies do not have pgf90 generate .d files so need to test
if the .d files are there first.

Improved gfortran dependency fix

Merge branch 'master' into scott/ksp-ksp-tutorials

Getting the gfortran -MMD bug working

3rd time is a charm: Only try to fix if there is a problem,
and remove just the bad lines.

Improve .d file generation

Crude fix for gfortran bug

gfortran produces bad .d files with the -MMD flag.
This creates a simple .d file that essentially links the .o file
to JUST the source file. Because this is just the test

Crude fix for gfortran bug

gfortran produces bad .d files with the -MMD flag.
This creates a simple .d file that essentially links the .o file
to JUST the source file. Because this is just the tests and just
f90, this seems like the most reasonable approach.

show more ...

Adding alt file replace functionality

Adding ALT=1 to the gmakefile.test arguments will enable passing the -M
to the shell harness which then passes it to petsc_diff which will then
replace the alt

Adding alt file replace functionality

Adding ALT=1 to the gmakefile.test arguments will enable passing the -M
to the shell harness which then passes it to petsc_diff which will then
replace the alt file. Should use with caution if multiple alt files are
present.

A bug was found while enabling this for alt files. I used a || operator
instead of a && operator.

show more ...

Merge branch 'jed/tests-cxx-linker'