ppc64 clang LNT buildbot / reference_output / test failures

Hal Finkel hfinkel at anl.gov
Tue May 14 16:27:03 PDT 2013


Renato,

I remember :) -- Thanks!

 -Hal

----- Original Message -----
> 
> Hi Hal,
> 
> 
> Remember the mismatch in ClamAV output you found locally? I also
> found here on my A9 box. Though, it doesn't show on my i7 laptop nor
> on my Chromebook.
> 
> 
> There's some random-hunting afoot. Now that the release 3.3 RC1 is
> behind, I'll have a look and see if I can spot the source of the
> randomness and hopefully fix it.
> 
> 
> cheers,
> --renato
> 
> 
> 
> On 8 May 2013 18:15, Hal Finkel < hfinkel at anl.gov > wrote:
> 
> 
> 
> 
> ----- Original Message -----
> > 
> > On 8 May 2013 17:16, Hal Finkel < hfinkel at anl.gov > wrote:
> > 
> > 
> > 
> > 
> > So these 'Matched signature' lines are appearing out of order
> > somehow. Will put the diff -u here:
> > http://fpaste.org/10881/79449211/
> > 
> > Do you have any idea what could cause this?
> > 
> > 
> > 
> > Hi Hal,
> > 
> > 
> > This looks very weird, I haven't seen that one yet. I thought it
> > was
> > just random position of the space, but it's not, and it's very
> > consistent.
> > 
> > 
> > It looks like a bug/feature in the ClamAV itself where the calls to
> > that print (or the function that prints it) depends on the platform
> > endianness or word size.
> 
> I'll double-check, but the problem is that it seems dependent on
> something else too. I don't see this difference on my PowerPC test
> system, and talking to Will on IRC, he said that he also does not
> see it when he runs the test suite from the command line, only from
> LNT.
> 
> -Hal
> 
> > 
> > 
> > cheers,
> > --renato
> 
> 



More information about the llvm-commits mailing list