fuel-dev team mailing list archive

[Dmitry Borodaenko <dborodaenko@xxxxxxxxxxxx>]

On Tue, Dec 17, 2013 at 3:33 AM, Mike Scherbakov
<mscherbakov@xxxxxxxxxxxx> wrote:
> Regarding docs: as Ivan mentioned, we have nailgun Python generated docs
> from there.
> If we move dev docs to fuel-docs, it will be impossible to generate our
> Nailgun Python documentation.

We can do the same thing we do in fuel-main (git clone --depth 1). In
fact, it makes sense moving the Python docs generation from fuel-web
to fuel-main and making generation of API docs for all of our
components a part of ISO build process. Benefits are many:

1) We can extend this to generate API docs for all components, not just Nailgun.
2) We already have to clone all components from separate Git
repositories during ISO build, so that steps is not duplicated.
3) We would be able to include a "guaranteed to be up to date" API
documentation with every ISO we build.

The downsides:

1) More work than leaving things as it is.
2) More build-dependencies for making an ISO.
3) More time to build an ISO from scratch.

> On Tue, Dec 17, 2013 at 3:11 PM, Ivan Kolodyazhny <ikolodyazhny@xxxxxxxxxxxx> wrote:
>> I'm partially agree with you. You touched more general problem, that
>> nailgun docs. Actually, there are a lot not web-related code in a fuel-web
>> repo. E.g. it will be good to move out at least fuel menu and fuel client to
>> separate repositories.

I think we should be careful about creating more repositories. We
already have many cases where we need to create separate review
requests for fuel-web and fuel-library for interdependent parts of the
same change, and one ends up being merged without the other. Not to
mention that each repository adds to operational overhead.
Conceptually, fuel menu and fuel client are all parts of UI for fuel,
it would be easier to keep them consistent if they're all part of a
single repository.

-- 
Dmitry Borodaenko