Fix WaitGroup usage

[rhcarvalho]

Calling wg.Done() does not "cancel" the WaitGroup, it does introduce a
bug in which depending on timing the WaitGroup counter might go
negative, what should never happen.

Would be nice to rewrite the method to be correct concurrent code, but
for now should be enough to avoid the panic condition and note that
there is a potential leak.

---

Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1370265

[rhcarvalho]

I propose this instead of #577, @gabemontero @bparees PTAL.
Sorry for coming with a PR, was easier to explain with code than lots of comments (without the code).

[rhcarvalho]

We don't need to initialize WaitGroups, just declare them and they are ready for use.
Idiomatic example: https://golang.org/pkg/sync/#example_WaitGroup

[gabemontero]

don't let the PM team see this :-) ... they prefer the :=

[bparees]

i also prefer := :)

[rhcarvalho]

This is not a necessary change, so will revert it to minimize the patch ;)

[gabemontero]

LGTM ... I'll delete the other pull when this merges.

@bparees - shall we assign the associated bugzilla and UPSTREAM vendor update to @rhcarvalho :-) ?

[rhcarvalho]

@bparees - shall we assign the associated bugzilla and UPSTREAM vendor update to @rhcarvalho :-) ?

Would be fun, but I don't want to overcommit :)

I'll be happy to send more s2i PRs when I come back (soon) 😄 , missing you guys

[bparees]

lgtm [merge]
@gabemontero please do the upstream PR to origin, we've already abused @rhcarvalho enough :)

[openshift-bot]

Evaluated for source to image merge up to b5d4d3e

[openshift-bot]

Source To Image Merge Results: SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pr_sti/207/)

[gabemontero]

roger that @bparees

[gabemontero]

and thx @rhcarvalho :-)