pcp
[Top] [All Lists]

pcp updates - pmUnitsStr_r fix and timezone shift strangeness in QAland.

To: pcp@xxxxxxxxxxx
Subject: pcp updates - pmUnitsStr_r fix and timezone shift strangeness in QAland.
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Tue, 07 Oct 2014 16:34:55 +1100
Delivered-to: pcp@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.1.2
Changes committed to git://git.performancecopilot.org/kenj/pcp.git dev

 man/man3/pmunitsstr.3  |    9 +++++++++
 qa/151                 |   13 +++++++++++++
 qa/746                 |   29 +++++++++++++++++++++++++++++
 qa/746.out             |   14 ++++++++++++++
 qa/group               |    1 +
 src/libpcp/src/units.c |    7 +++++++
 6 files changed, 73 insertions(+)

commit d57df37c5c8a83fde57476592b73d51bd62e8092
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Tue Oct 7 16:28:29 2014 +1100

    libpcp - fix for pmUnitsStr_r buffer overrun
    
    Since the definition of pmUnitsStr_r() dictates that buflen be at least
    60, enforce this on entry.  There is a small ABI change here that in
    the event that buflen is too short, pmUnitsStr_r will now return NULL
    instead of buf.
    
    Updated the man page.
    
    Added qa/746 to tickle this buglet and prove it is fixed.
    
    Addresses http://oss.sgi.com/bugzilla/show_bug.cgi?id=1068

commit 0eb6867db4eabf5774b6324da3f7e1748b335436
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Tue Oct 7 12:09:05 2014 +1100

    qa/151 - don't run for 6 days after a timezone change
    
    Strange but true!  pmlogger_daily is doing the right thing, but
    this QA test is confused ... simply skip it until things settle
    down (the 6 days comes from the test, it is not make believe).

<Prev in Thread] Current Thread [Next in Thread>