g+
g+ Communities
Argonne National Laboratory

Experimental Physics and
Industrial Control System

2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  <20122013  2014  Index 2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  <20122013  2014 
<== Date ==> <== Thread ==>

Subject: Re: Base R3.14.12.3-pre1 available
From: <nick.rees@diamond.ac.uk>
To: <anj@aps.anl.gov>
Cc: core-talk@aps.anl.gov, Franck.DiMaio@iter.org
Date: Fri, 30 Nov 2012 20:55:07 +0000
Hi Andrew,

Thanks. Regarding the working TCP stack on Windows the answer is  (I think) yes - it is our Windows terminal server machine and I was logged in using remote desktop at the time. However that doesn't rule out other problems. The compiler and c support libraries are also fairly recent. Any diagnostic suggestions would be welcome. I will also run the R3.14.12.1 tests to see if the result is the same.

Cheers,

Nick Rees
Principal Software Engineer           Phone: +44 (0)1235-778430
Diamond Light Source                  Fax:   +44 (0)1235-446713

Andrew Johnson <anj@aps.anl.gov> wrote:
Hi Nick,

On 2012-11-30 nick.rees@diamond.ac.uk wrote:
> I have just tried building R3.14.12.3-pre1 on Windows server 2008R2 and get
>  the following error - and I am not enough of a C++ expert to unwind what
>  is needed:

Thanks for that report, Janet confirmed it so I just reverted the commit that
modified epicsSingleton.h a couple of weeks ago.

> I also tried building on RHEL6, linux-x86_64. This worked.

[I usually use 'make -s runtests' to silence the output from GNUmake and just
display the test results BTW.]

> The test output
> is:
>
> [npr78@pc0018 test]$ make runtests
...
> All tests successful.
> Files=25, Tests=2006, 396 wallclock secs ( 0.48 usr  0.08 sys +  1.01 cusr
>  0.58 csys =  2.15 CPU) Result: PASS

> Is this what is expected? How many tests are meant to pass?

Those numbers are right for src/libCom/test; if you do the 'make runtests'
from the top level directory it also runs a second set of tests in src/db/test
but they're not likely to fail.

> Finally, because I couldn't test R3.14.12.3, I tried to run make runtests
>  on a fully patched version of R3.14.12.2 on Windows Server 2008R2 64 bit
>  and got:

> epicsCalcTest.t ............ Failed 2/566 subtests
> epicsMathTest.t ............ Failed 6/35 subtests

Those failures are known bugs in some versions of Microsoft's optimizer, which
assumes for example that the expression X-X will always give 0 (not true when
X is NaN).

> blockingSockTest.t .........
>
> A call to 'assert(this->sock != INVALID_SOCKET)'
>     by thread 'win2a8' failed in ..\blockingSockTest.cpp line 172.
> EPICS Release EPICS R3.14.12.2 $Date: Mon 2011-12-12 14:09:32 -0600$.
> Local time is 2012-11-30 18:30:58.382980961 GMT Standard Time
> Please E-mail this message to the author or to tech-talk@aps.anl.gov
> Calling epicsThreadSuspendSelf()

Interesting, we don't get that here; that's one of Jeff's tests, and it
implies that socket() returned an invalid socket.  Did you have a working
TCP/IP network connected to the machine at the time?

Thanks,

- Andrew
--
Computer science is as much about computers as astronomy is about
telescopes. -- Edsger Dijkstra


Navigate by Date:
Prev: Re: Base R3.14.12.3-pre1 available Andrew Johnson
Next: RE: Base R3.14.12.3-pre1 available nick.rees
Index: 2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  <20122013  2014 
Navigate by Thread:
Prev: Re: Base R3.14.12.3-pre1 available Andrew Johnson
Next: [Bug 1083193] Re: Lifetime of CALLBACK undocumented Ralph Lange
Index: 2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  <20122013  2014 
ANJ, 03 Dec 2012 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· EPICSv4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·