alltests: fix to correctly test a prefix install as per 8e69c5ec7dbec5c4ff0e2255fa41e86223635a8c

CUSP: Removed three left-overs from bfort-petsc.txt.

Reported-by: Jose Roman <jroman@dsic.upv.es>

manimplementations: print current directory, to allow more output in docs logs

Docs: implementations scraping: add PETSC_INTERN functions (e.g. DMClone_Plex)

Merge branch 'jed/faster-pipelines' [PR #908]

* jed/faster-pipelines:
petsc_testrun: remove watchtime.sh in favor of ulimit -St
petsc_testrun: use exit code even when no job control
make: use

Merge branch 'jed/faster-pipelines' [PR #908]

* jed/faster-pipelines:
petsc_testrun: remove watchtime.sh in favor of ulimit -St
petsc_testrun: use exit code even when no job control
make: use and recommend "make check" instead of "make test" for the brief check
config make: MAKE_TEST_NP default to MAKE_NP for with-mpi=0
travis: update (deactivated) alltests for consistency with bitbucket-pipelines
pipelines: use COPTFLAGS=-O
pipelines: use -check_pointer_intensity 0 -malloc_test

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 ...

make: use and recommend "make check" instead of "make test" for the brief check

This is a more standard name and thus we can endorse a terminology where
"test" means from the full suite (using the g

make: use and recommend "make check" instead of "make test" for the brief check

This is a more standard name and thus we can endorse a terminology where
"test" means from the full suite (using the gmake-based system).

show more ...

Merge remote-tracking branch 'origin/balay/remove-old-regression'

makefile: add + to recursive recipes so that "make test" does not warn about -j

pipeline: trap more failures.
travis-ci: update build to use all available cores

build: remove old 'dos' related build files - which have not been used in ages

bitbucket: pipeline: update error check

delete files not used anymore

build: ALLTESTS_PRINT_PROGRESS flag doesn't exist anymore - so remove its reference

Merge branch 'maint'

install: cannot use PREFIXDIR as a flag to determine prefix install anymore. Use text from 'shared_install' target in petscrules

gcov: PETSC_ARCH is required onlyfor -run_gcov - and not for -merge_gcov [where it doesn't exist]

build: remove old test targets from makeifle

Also remove TESTMODE code from buildtest

dashboard: if unable to parse timestamps in logs - return 0 for timediff [and not crash/burn]

dashboard: date -R is not portable :( [ aka 10.11.6 petsc-mini]. So explicitly specify format

dashboard: use consistant date format across builds. i.e rfc-email/RFC-5322

runhtml: fix to handle missing TIMZONE in logs [aka cygwin]. perhaps we should use 'date -R' on all builds

dashboard: if logs have (start) timestamp older than 12h - then mark them as stale.

Fix date in logs to include timezone so as to avoid errors of type:
TypeError: can't subtract offset-naive and off

dashboard: if logs have (start) timestamp older than 12h - then mark them as stale.

Fix date in logs to include timezone so as to avoid errors of type:
TypeError: can't subtract offset-naive and offset-aware datetimes

show more ...

runhtml: replace manual code to diff time with dateutil.parser code

dashboard: fix logs to have consistant format for date/time so its easier to parse