Open Stack

On Wed, May 11, 2016, at 01:11 PM, Robert Collins wrote:
> So, given that that is the model - why is language part of it? Yes,
> there are minimum overheads to having a given language in CI - we need
> to be able to do robust reliable builds [or accept periodic downtime
> when the internet is not cooperating], and that sets a lower fixed
> cost, varying per language. Right now we support Python, Java,
> Javascript, Ruby in CI (as I understand it - infra focused folk please
> jump in here :)).

The actual list is more accurately Python and Bash and to a degree
Javascript. We can run Java and Ruby and there have even been people
running Go, this is a result of our test infrastructure giving you root
on single use test instances. But we haven't really jumped over
backwards to properly support these languages with package mirrors,
debugging help, test framework help, etc this has largely only been done
for Python and Bash (but more often Javascript too due to the realities
of the modern browser world).

One thing to keep in mind is that people working cross project tasks and
particularly infra and gate triage end up doing a significant amount of
debug work keeping the projects running. Whether it is an upstream lib
breaking, packaging tools exploding, packages disappearing, or
misdiagnosed infrastructure issues these individuals do spend a ton of
time knee deep in language specific details. Yes, there is a minimum
overhead for supporting a given language, and that overhead is much
higher than many seem to realize. This is why language is part of it.

If we had a stronger culture of cross project cooperation and OpenStack
issue triage I definitely wouldn't worry about this as much, but the
reality of the current world is that a very small subsection of our
community is involved in this work and they already have a high rate of
burn out.

I am glad that general consensus seems to be we should have this
discussion and determine these costs before we make any decisions.

Clark