Re: Two glitches while testing gcc version 4.3.0 20070216

[Brooks Moses <brooks dot moses at codesourcery dot com>]

Steve Kargl wrote:
> On Sat, Feb 17, 2007 at 06:33:53PM +0100, Dominique Dhumieres wrote:
> > While testing gcc version 4.3.0 20070216 on OSX 10.3.9, I saw the following i
> > problems:
> > (1) regtesting gfortran as an extra failure:
> >
> > ...
> > Setting LD_LIBRARY_PATH to .:.:.:.
> > FAIL: gfortran.dg/secnds.f  -O3 -fomit-frame-pointer -funroll-all-loops 
> > -finline-functions  execution test
> > Executing on host: /sw/lib/gcc4//bin/gfortran /Users/dominiq/test
> > /gcc-4.3-20070216/gcc/testsuite/gfortran.dg/secnds.f   -O3 -g  -O0 
> > -ffloat-store   -lm   -o ./secnds.exe    (timeout = 300)
> > ...
> >
> > Is this related to 
> >
> > [patch, fortran, testsuite] Replace secnds*.f 20-ms tolerance with bracketing
> >
> > or is this a new problem?
>
> It's simply a timing problem.  secnds.exe is trying to make a timing
> measurement and the tolerance is IMHO too strict.

Theoretically, though, the patch he's referring to (which I put in) 
should have fixed that problem, by eliminating tolerances -- it calls 
SECNDS, calls DATEANDTIME, and calls SECNDS again, and checks that the 
DATEANDTIME value is between the two SECNDS values.  So it shouldn't be 
doing that any more.

However, after submitting the patch, I did discover with my 
"instrumented" version of the testcase that it will very occasionally 
still fail; the intermediate DATEANDTIME result was 40ms ahead of the 
first SECNDS result, but the SECNDS result was only 20ms ahead.  Since 
it's never happened again for me, I haven't looked closer to see if 
there's a "real" reason for it, but my guess is that it's a rounding issue.

I trust that this isn't happening repeatably, Dominique?  If you re-run 
that testcase a few times, does it keep failing?
 (make check-fortran RUNTESTFLAGS="dg.exp=gfortran.dg/secnds.f90")

Maybe I should put a loop into the testcase to run it two or three times 
and only abort if all of the passes fail.

- Brooks