[ofiwg] Travis CI testing of libfabric and fabtests

[Jeff Squyres (jsquyres)]

On Oct 12, 2015, at 12:21 PM, Hefty, Sean <sean.hefty at intel.com> wrote:
> 
>> FWIW, we've left such details to the sites running the CI tests.
> 
> But in this case, the site I'm referring to is github - not Cisco or Intel testing.  

I'm not sure I understand the distinction...?

> The Travis CI testing is happening automatically for each pull request.  I don't know what physical site is doing the actual testing, or even what company is paying for the costs.  From my perspective, it's just a wonderful free service for open source projects.  :)
> 
> But for the build testing to work, we need to have the .travis.yml script pull in something. 

I can't speak for Travis, but for our Jenkins testing (which we still need to hook up to the ofiwg Github PR results...), we just have this in the Jenkins build script:

./configure                 \
    --prefix="$prefix"      \
    --enable-usnic=dl       \
    --enable-sockets=dl     \
    --disable-verbs         \
    2>&1 | tee c.txt

This ensures that the usnic provider is built (and errors if it doesn't).

There are no specific dependent libraries needed here, but in a similar Open MPI CI build, we have configure lines like this:

./configure --with-hwloc=/home/mpiteam/hwloc-1.8.1/install

just to pick up a specific version of hwloc.

All I'm saying is that we don't all have to test exactly the same thing (i.e., use the exact same versions of dependent libraries).  We just want the union of everything tested to pass.  When something fails, we figure out how to fix it -- which may be (for example), "Hey ABC, please update your dependent library XYZ to vJ.K.L."

-- 
Jeff Squyres
jsquyres at cisco.com
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/