[petsc-dev] next tests, absoft and ifc

[Barry Smith]

> On Apr 13, 2017, at 8:38 PM, Tobin Isaac <tisaac at uchicago.edu> wrote:
> 
> On Thu, Apr 13, 2017 at 04:51:23PM -0600, Scott Kruger wrote:
>> 
>> 
>> This was fixed awhile ago and folded into the altfiles pull request, and
>> even at one point had a cherry-picked merge into one of Matt's pull requests
>> (it's a single line in config/testparse.py).
> 
> There is a different solution to that problem already in next from
> a12800a.  I do not think this is the same problem: a problem in a
> python script would manifest in all builds on the same machine, but my
> problem is limited to specific builds.
> 
> At any rate, I'm giving up, accepting that computers are stupid, and
> replacing "::ascii_*" with "ascii::ascii_*" so that I don't have to
> wait for this to be diagnosed and fixed before cleaning up the nightly
> tests.
> 

  Don't give up! 

  So you think it is the C coding handling of ::something arguments? on specific machines?

  Satish can track that down easily on those machines.


>> 
>> I am catching up after some vacation time and don't know the status of the
>> altfiles merge.  I thought I had resolved all of the issues Satish raised,
>> but it's still open so there must be something...
>> 
>> Scott
>> 
>> 
>> On 4/13/17 10:29 AM, Tobin Isaac wrote:
>>> 
>>> Hi Satish,
>>> 
>>> I'm working to get all of our dm tests to pass in next.  We have a lot
>>> of output that comes from "::ascii_info_detail" being passed to
>>> DMViewFromOptions().  On two builds, linux-gcc-absoft_churn and
>>> linux-gcc-ifc-cmplx_crank, it looks the viewer format is being ignored
>>> and short output is being generated instead.
>>> 
>>> I can't reproduce this behavior on my machine, but I notice that there
>>> are no diff errors in the few instances where
>>> "ascii::ascii_info_detail" is given instead.
>>> 
>>> Could you please run on those two next builds
>>> 
>>>    PETSC_OPTIONS="-options_view" make -f gmakefile dm_impls_plex_tests-runex5_0
>>> 
>>> to see if somehow keys starting with colons are not being parsed
>>> correctly?
>>> 
>>> Thanks,
>>>  Toby
>>> 
>> 
>> -- 
>> Tech-X Corporation               kruger at txcorp.com
>> 5621 Arapahoe Ave, Suite A       Phone: (720) 974-1841
>> Boulder, CO 80303                Fax:   (303) 448-7756