pcp
[Top] [All Lists]

pcp updates - assorted

To: pcp@xxxxxxxxxxx
Subject: pcp updates - assorted
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon, 10 Nov 2014 09:40:14 +1100
Delivered-to: pcp@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.2.0
Changes committed to git://git.performancecopilot.org/kenj/pcp.git dev

 qa/003                 |    1 
 qa/752.out             |  228 ++++++++++++++++++++++++-------------------------
 qa/944                 |    1 
 qa/src/rtimetest.c     |   90 ++++++++++++++++++-
 src/include/buildrules |    2 
 5 files changed, 203 insertions(+), 119 deletions(-)

commit d1ab4c41b9be162a3245a1d350bfd5cf2e180f9d
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Mon Nov 10 09:39:19 2014 +1100

    qa/src/rtimetest.c: improve diags, accept command line args
    
    On the trail of the "Next Tuesday" bug ...

commit ecc90bf9f0b23e9bc1a2445cb8eacd5638df26d7
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Mon Nov 10 09:37:24 2014 +1100

    buildrules: need rm -f before ln -s
    
    In the build this was hidden because the alternate named python and
    perl files did not exist.
    
    But the same rule is used in qa/src and there the targets may exist
    and it was only a matter of time before this bit somone ... like me
    a couple of days ago.

commit 5e32d07b38272fda1bce37f24449b352cec9eac7
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Fri Nov 7 06:46:52 2014 +1100

    qa/752: tidy up output from rtimetest.c
    
    Previous commit made the formatting harder to read which was the exact
    opposite of the intention.
    
    The "Tuesday" bug remains unresolved at this point in time.

commit 9b736c050261be3ae9abff857004449a3cd63659
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Fri Nov 7 06:36:18 2014 +1100

    qa/003: hinv.node.online not always available, add to filter

commit db2bd86aa051f8c7d47836002b7d2ab1d291c930
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Thu Nov 6 09:48:56 2014 +1100

    qa/944: better handling of multiple users being in the same default group
    
    For example on grundy, need to handle case when the target is _not_ first
    in the list, ... so filter
     12(games),1000(kenj),1001(trev)
    to produce
     GROUPID(GROUPNAME),...
    not
     12(games), GROUPID(GROUPNAME),...

<Prev in Thread] Current Thread [Next in Thread>
  • pcp updates - assorted, Ken McDonell <=