pcp
[Top] [All Lists]

pcp updates

To: Nathan Scott <nscott@xxxxxxxxxx>
Subject: pcp updates
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sat, 31 Jan 2009 11:51:56 +1100
Cc: pcp@xxxxxxxxxxx
In-reply-to: <1233311014.4566.11.camel@xxxxxxxxxxxxxxxxxx>
References: <1233203077.4580.30.camel@xxxxxxxxxxxxxxxxxx> <1233299842.9934.40.camel@xxxxxxxxxxxxxxxxxxxxxxx> <1233311014.4566.11.camel@xxxxxxxxxxxxxxxxxx>
Reply-to: kenj@xxxxxxxxxxxxxxxx
On Fri, 2009-01-30 at 21:23 +1100, Nathan Scott wrote:
> On Fri, 2009-01-30 at 18:17 +1100, Ken McDonell wrote:
> > Not a quirk at all -- a bug.  Any aggregate operator (as the top
> > left-most node in the expression tree) returning a scalar value would
> > suffer the same way.
> > 
> > I have a fix that I'll push once it passes all the rest of the pcp qa.

Fix now passes QA, so here it is.

Changes committed to git://oss.sgi.com/kenj/pcp.git dev

 src/pmie/src/show.c |   19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

commit 7688dded9fc9d82a9a4ae90b2f8b854b947f16b2
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date:   Sat Jan 31 11:43:58 2009 +1100

    Fix %v value extraction for aggregate operators.
    
    Problem found by Nathan Scott with this sort of expression
        test2 = sum_inst(kernel.all.load) >= (kernel.all.load #'1 minute')
        -> print "TEST2: Left-most sum_inst: v=%v";
    Code incorrectly assumed top left-hand expression in the tree had type
    boolean AND then descended below the aggregate to try and pick from the
    set of values over which the aggregate was evaluated.



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