Changes committed to git://git.pcp.io/kenj/pcp master
Ken McDonell (7):
qa/578: still dinkin' with acceptable openfds range
qa/580: filter to catch yet another "init" process name
qa/340: general cleanup
qa/666 & qa/669: wait for pmlogger in cleanup()
qa/381: more armour, more diagnostics
qa/808: be a bit more patient and wait for processes to really exit
qa/720: valgrind is broken on Ubuntu 16.04 ... just notrun here
qa/.gitignore | 1
qa/340 | 141 ++++++++++++--------------------------------------------
qa/340-v1v2.out | 10 ---
qa/340-v2.out | 36 --------------
qa/340.out | 34 +++++++++++++
qa/381 | 61 +++++++++++++++---------
qa/578 | 6 +-
qa/580 | 1
qa/666 | 1
qa/669 | 1
qa/720 | 6 ++
qa/808 | 21 ++++++++
12 files changed, 139 insertions(+), 180 deletions(-)
Details ...
commit 253575c6f44095358cffe1f3fd9a4cfd88a63c42
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Jul 25 12:53:40 2016 +1000
qa/720: valgrind is broken on Ubuntu 16.04 ... just notrun here
commit bbf9e93683e5603c64c9066d87f18e7d09a28727
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Jul 25 12:53:00 2016 +1000
qa/808: be a bit more patient and wait for processes to really exit
commit aaee00cd07a293c6a04eaa533e4bc9625ca59c46
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Jul 25 12:51:57 2016 +1000
qa/381: more armour, more diagnostics
This one is proving very difficult to get to pass 100% of the time.
commit 6757a5167ea4d5040e525d7bc74f3932291812b3
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Jul 25 12:49:56 2016 +1000
qa/666 & qa/669: wait for pmlogger in cleanup()
The new 870 checker was tripping over this if pmlogger had not
completely restarted before the 66[69] test exited ... in theory this
could have caused a cascaded error for the next QA script to be run.
commit 30d5c29c7e6a2e6a626132bef913a0550ca082e0
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Jul 25 07:10:51 2016 +1000
qa/340: general cleanup
1. remove tests for v1 pmcd protocols ... not sure anyone even
even remembers the pre-distributed PMNS days
2. remove the variant output files
3. clean up the "is pmcd alive" probing to use -h unix: in preference
to -h localhost: where possible ... this seems to make the test more
likely to pass on some platforms.
commit 2cc9d56abcad5ed56a0d5e34921125be7bd63ce2
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon Jul 25 07:07:06 2016 +1000
qa/580: filter to catch yet another "init" process name
On Ubuuntu 16.04 (bozo) it is /lib/systemd/systemd
commit 6dc5d93fddc694aa069e129f7f5f440d29122efc
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun Jul 24 21:42:12 2016 +1000
qa/578: still dinkin' with acceptable openfds range
|