Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Include stdbool.h in pbs_ifl.h #426

Open
wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Include stdbool.h in pbs_ifl.h #426

wants to merge 2 commits into from

Conversation

tabaer
Copy link
Contributor

@tabaer tabaer commented May 24, 2017

Needed to build things like pbs-drmaa (http://apps.man.poznan.pl/trac/pbs-drmaa)

@tabaer
Copy link
Contributor Author

tabaer commented May 25, 2017

I submitted this pull request yesterday, and its Travis-CI test builds reported failures in the mail_throttler and machine tests. I have attempted to reproduce that on a CentOS 7.2 test VM and ran into some difficulties determining what the problems are:

  1. After doing a sucessful "./configure --prefix=/usr/local/torque-git --enable-maintainer-mode --with-check --with-coverage ; make -j4" on my test system, several of the tests run by "make check" (specifically machine, numa_chip, numa_core, numa_pci_device, and numa_socket) have code that attempts to include hwloc.h even if neither cpusets nor cgroups were enable at configure time, and those failed initally because I did not have hwloc-devel installed at the time. This was easy enough to work around by installing hwloc-devel, but it seems curious to me for tests to depend on libraries and header files that are not needed by the configured build. Shouldn't the NUMA tests be skipped unless either cpusets or cgroups are enabled?

  2. My build also had the same failure in the mail_throttler tests as in the Travis-CI build, but the log was not particualrly helpful:

[root@torque-drmaa torque-git]# cat src/test/mail_throttler/test-suite.log
[...]
Running suite(s): mail_throttler test suite methods
50%: Checks: 2, Failures: 0, Errors: 1
test_uut.c:10:E:test_adding:test_adding:0: (after this point) Test timeout expired

This is not exactly an unambiguous failure mode. It also seems curious that conditionally including stdbool.h in pbs_ifl.h would not break the compilation of the software but would somehow break the mail_throttler unit tests.

  1. Unlike the Travis-CI builds, my build did not fail in the machine tests. All four Travis-CI builds failed as follows:

make -C machine check
make[3]: Entering directory /home/travis/build/adaptivecomputing/torque/src/test/machine' CXX machine.lo CXX jsoncpp.lo CXX numa_constants.lo CXXLD libuut.la CC scaffolding.lo CCLD libscaffolding.la make test_uut ../coverage_run.sh make[4]: Entering directory /home/travis/build/adaptivecomputing/torque/src/test/machine'
CC test_uut.o
CCLD test_uut
make[4]: Nothing to be done for ../coverage_run.sh'. make[4]: Leaving directory /home/travis/build/adaptivecomputing/torque/src/test/machine'
make check-TESTS
make[4]: Entering directory `/home/travis/build/adaptivecomputing/torque/src/test/machine'
Running suite(s): machine test suite methods
80%: Checks: 10, Failures: 2, Errors: 0
test_uut.c:109:F:test_place_and_free_job:test_spread_place:0: called 0
test_uut.c:84:F:test_place_and_free_job:test_place_all_execution_slots:0: placed all: 0
FAIL: test_uut
File: machine
�[1;33mTOTALCOV -- machine.cpp: Lines(503) - executed:50.50%�[0m
PASS: ../coverage_run.sh

1 of 2 tests failed
Please report to [email protected]

Since my local builds of this do not have this problem, I have no idea how I could fix it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant