WIP: Migrate to dep

[aledbf]

• update development.md
• add note about the go-bindata module (only remaining git module)
• fix federation build

[chrislovecnm]

/assign @justinsb

I know you reached out to him on this one

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
We suggest the following additional approver: justinsb

Assign the PR to them by writing /assign @justinsb in a comment when ready.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these OWNERS Files:

• OWNERS

You can indicate your approval by writing /approve in a comment
You can cancel your approval by writing /approve cancel in a comment

[aledbf]

@justinsb ping

[chrislovecnm]

Is it actually working? I am impressed. I will test when I can.

[chrislovecnm]

How are we going to not run into problems like this?

I am only trying to add aws go sdk to test infra, and they are using bazel and dep together. kubernetes/test-infra#6086 - it is a mess.

I am wondering if we need to wait for go_rules to mature a bit with dep.

Thoughts?

[aledbf]

How are we going to not run into problems like this?

I don't know.
That being said I can commit (long term) to fix any issue with dep in this repository.

I am only trying to add aws go sdk to test infra, and they are using bazel and dep together. kubernetes/test-infra#6086 - it is a mess.

I cannot event use the current code from master

aledbf@me:~/go/src/k8s.io/test-infra$ planter/planter.sh bazel build //...
planter/planter.sh: 58: planter/planter.sh: UID: parameter not set

I am wondering if we need to wait for go_rules to mature a bit with dep.

Right, but removing 6 millions of lines from this repository and almost every git module is a huge win (personal opinion)

[chrislovecnm]

First off I appreciate the effort and how you have helped this project. Huge kudos on that.

Dep is a huge win if we can still use bazel, otherwise let sleeping dogs lie until bazel and dep start playing nicer together. I am not a fan of the submodules, but it works, and is stable.

Having bazel broken is a HUGE loss for the project. So if we can get them working together nicely I would be open to this. Touching out build system makes me a tad nervous. We still have some makefile issues fr the last set of changes.

Again appreciate the time an effort you took with this.

[aledbf]

Dep is a huge win if we can still use bazel, otherwise let sleeping dogs lie until bazel and dep start playing nicer together. I am not a fan of the submodules, but it works, and is stable.

Which tasks require bazel? (just to make sure it works)

[chrislovecnm]

Require bazel none. We do have bazel build and test targets. Admittedly it is the work flow that I am nervous about ... Investigating in the #bazel channel as well.

@justinsb thoughts?

[aledbf]

@chrislovecnm as this is a WIP 😉 the last commit starts the process to fix the dep <-> bazel issues

make bazel-gazelle
gazelle update-repos -from_file Gopkg.lock
gazelle update -external vendored
make bazel-build

[justinsb]

That looks better than the approach I've been using - I'll have to check out your trick @aledbf

One thing we want to be sure of is that the versions don't change - I created a script hack/dep.py which cross-checks dep status against the k8s Godeps. I used the k8s release-1.9 branch and there were a lot of diffs. I think we want to lock almost of all our dependencies, which in practice means we're not really using dep package resolution - but that is not necessarily a bad thing!

That script is #4152

[chrislovecnm]

Talking with the infra-test folks dep has some bugs, especially with prune.

I think we should discuss this on next office hours.

[k8s-github-robot]

@aledbf PR needs rebase

[chrislovecnm]

@aledbf / @justinsb we did not discuss this during office hours, meant to. This PR is going to be in rebase hell shortly. I am kinda on the fence about this one, but I will trust your guy's judgment.

What are the final thoughts on this? Go or no GO?

[aledbf]

What are the final thoughts on this? Go or no GO?

I take this as an experiment. Not sure if make sense to merge this PR.
Can we leave this open just as a WIP until I can solve the bazel - dep issues?

[justinsb]

Thanks for getting us to do this @aledbf :-)