No worries! Advertising for interest is still good at this point.
Refining tasks to be an appropriate length seems prudent now.
WRT infra, it depends a lot on what you (collective "you" as the
project) want to see tested and how you as developers get the test
results. For example:
* Do devs deploy on their own machines (local, ec2) or use some
packaging (Vagrant, Docker, RPMs/DEBs+scripting)
* What defines "correctness" for Rya? What about "performance"? How do
you measure them?
* How far do unit tests get you? Integration tests? What, if any, new
tests does Rya need?
At a high level, that's what is always on my mind. I don't have enough
context to say what I think Rya should have, but that's at least my
thought process :)
Adina Crainiceanu wrote:
Josh, I completely agree with all your comments, and thank you for taking
the time to provide input.
Would it be possible for you to elaborate on what needs to be implemented
to have the reliable testing infrastructure you mentioned?
I realize that I should have had a better plan before proposing the
continuous integration project for GSoC, but now it is out there and a
student is interested in it. I hope that you and other people on this list
can help with fleshing out a plan for implementing the testing
infrastructure, so the project is fit for GSoC. I am very happy to
supervise the student if the project/student get selected for GSoC.
Thank you very much,
Adina
On Sat, Feb 20, 2016 at 3:00 PM, Josh Elser<josh.el...@gmail.com> wrote:
Well to avoid my original reaction being taken out of my intended context,
an umbrella testing/integration gsoc project would be cool. I just meant
that making Jenkins run sounds like a one week effort at best.
Getting some reliable testing infrastructure, especially with all the
databases that Rya supports would be super cool.
On Feb 19, 2016 8:56 AM, "Adina Crainiceanu"<ad...@usna.edu> wrote:
I was looking at
http://en.flossmanuals.net/GSoCMentoring/defining-a-project/ for project
ideas and infrastructure/automation was one of the ideas, but I agree
that
it is not the typical project and might not be the right scope/flavor.
Should I remove the gsoc2016 label so it does not show up in the list?
I also listed the search over multiple Rya instances as one of the
projects. I think a simple version can be implemented by a good student
in
3 month, and that is more inline with the other projects.
On Fri, Feb 19, 2016 at 11:44 AM, Josh Elser<josh.el...@gmail.com>
wrote:
re Jenkins server, yes. https://builds.apache.org
Assuming it hasn't changed since I was involved, GSoC should be a
programming task that a student could accomplish in ~3months with a
strong
weekly effort (I'm not sure if it's quantified in hrs/week).
Some examples from 2015:
https://www.google-melange.com/gsoc/project/details/google/gsoc2015/shuxiong/5796788510392320
https://www.google-melange.com/gsoc/project/details/google/gsoc2015/radu_manole/5668600916475904
https://www.google-melange.com/gsoc/project/details/google/gsoc2015/adperezmorales3/5717271485874176
I was pulling these from
https://www.google-melange.com/gsoc/projects/list/google/gsoc2015
Puja Valiyil wrote:
I don't think setting up some Jenkins jobs would be difficult.
There's
a
nice GUI for Jenkins that guides you through it and since we cleaned
up
the
poms recently it should be straight forward.
Does apache have a Jenkins server we could use?
Sent from my iPhone
On Feb 18, 2016, at 2:23 PM, Adina Crainiceanu<ad...@usna.edu>
wrote:
Hi all,
I though that one of the proposed projects for Rya at GSoC could be
to
have
a student working on setting up continuous integration / automatic
builds
for Rya. Those with more experience in something like that, can you
please
comment on what would be needed for such a project?
Thanks,
Adina
--
Dr. Adina Crainiceanu
http://www.usna.edu/Users/cs/adina/
--
Dr. Adina Crainiceanu
http://www.usna.edu/Users/cs/adina/
