pcp
[Top] [All Lists]

pcp updates - getting back in sync

To: pcp@xxxxxxxxxxx
Subject: pcp updates - getting back in sync
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri, 13 Feb 2015 20:30:51 +1100
Delivered-to: pcp@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.4.0
This is getting pretty messy ... we're doing a revert dance.  These changes 
applied to my git repo make mine the same as the "official" one plus the 
valgrind-suppress and pcp-push changes.

Nathan/Dave, it is probably going to need some conflict mangling at your end, 
and then I'll have to do some at my end when you commits come back down.

Sigh.

Changes committed to git://git.pcp.io/kenj/pcp dev

Ken McDonell (3):
      Revert "Revert "libpcp: coverity-detected leak return code and null check 
fixes""
      qa/valgrind-suppress: rework Frank's supressions for RHBZ1191183
      scripts/pcp-push: make take mail more verbose by default

Nathan Scott (3):
      qa: fix memcached output for bytes_read filtering
      Revert "libpcp: coverity-detected leak, return code and null check fixes"
      pmdalinux: improve /proc/cpuinfo parser for irregular formats

 qa/835.out                     |    1 
 qa/885.out                     |   89 ++++++++++++++++++++++++++++++-----------
 qa/valgrind-suppress           |   19 ++++++--
 scripts/pcp-push               |   15 ++++++
 src/libpcp/src/auxconnect.c    |    8 ++-
 src/libpcp/src/auxserver.c     |   20 ++++-----
 src/libpcp/src/getopt.c        |    6 ++
 src/libpcp/src/logportmap.c    |   12 ++---
 src/libpcp/src/spec.c          |    6 +-
 src/pmdas/linux/proc_cpuinfo.c |   57 ++++++++++++++------------
 src/pmdas/linux/proc_cpuinfo.h |    6 --
 11 files changed, 157 insertions(+), 82 deletions(-)

Details ...
commit 7ee069391db5d687a80524417c81d452adab1ae7
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Fri Feb 13 20:24:46 2015 +1100

    scripts/pcp-push: make take mail more verbose by default
    
    Old behaviour is with -s (new option).
    
    New default behaviour is to include commit logs with details of
    changes, not just one-line summaries and diffstats.

commit 5b0faa31bd98663d65e3a003600954921e3af808
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Fri Feb 13 20:23:48 2015 +1100

    qa/valgrind-suppress: rework Frank's supressions for RHBZ1191183

commit a6d6856792319a77811b8271211a718643d147be
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Fri Feb 13 20:14:56 2015 +1100

    Revert "Revert "libpcp: coverity-detected leak, return code and null check 
fixes""
    
    This reverts commit f1b793328b3eb134157091a2260ab7cfab28135a.
    
    Trying to get back in sync with Nathan's changes.

commit 05c9881334c74f8b1e9899e40b144ad0ad39778e
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date:   Fri Feb 13 17:30:23 2015 +1100

    pmdalinux: improve /proc/cpuinfo parser for irregular formats
    
    For a long time the architectures with /proc/cpuinfo forms that
    have global CPU information either before/after the individual
    "processor:" lines have been problematic.  The last wrinkle was
    on s390 where the global info precedes these lines (previously
    we'd only seen archs with processor lines at the files head).
    
    Rework things so that global information is saved on-stack in
    the same way for all platforms (removing Alpha cpp conditional
    code, and improving detection on Alpha at the same time - as
    can be seen in qa/885).  The test 885 expected output has been
    updated to include the new s390x output, which previously was
    causing a sigsegv.

commit 476df6bde7b5288786e8d996bf49fc70dcc8feaa
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date:   Fri Feb 13 17:24:34 2015 +1100

    Revert "libpcp: coverity-detected leak, return code and null check fixes"
    
    This reverts commit d3895954cc63aa59ed84aec6b98becfeaf0c230c.
    
    (problematic parts reworked and/or removed).

commit c18339170270b0677df47281ed84a552a4445b8b
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date:   Thu Feb 12 20:03:57 2015 +1100

    qa: fix memcached output for bytes_read filtering

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