use the global default socket path for spire-agent run

[joewilliams]

Pull Request check list

• Commit conforms to CONTRIBUTING.md?
• Proper tests/regressions included?
• Documentation updated?

Affected functionality
Default socket path for spire-agent run.

Description of change
Make the default spire-agent default socket path more obvious by using the global default value.

Which issue this PR fixes

[azdagron]

Code change is looking good!

There is a small risk of breaking backwards compatibility here for folks who don't define the socket_path in their agent configuration who were relying on the strange ./spire_api default. Considering all of the configurations defined by tutorials and examples do define socket_path, it doesn't seem too likely that someone would be reliant on this, but I thought I'd call out the risk.

Thoughts @evan2645 , @amartinezfayo , @APTy , @mcpherrinm? Any concern taking this potentially breaking change? What is your evaluation of the risk?

[joewilliams]

Not sure why this test is failing while the others are passing:

golangci/golangci-lint info installed /home/travis/gopath/src/github.com/spiffe/spire/.build/linux-x86_64/golangci_lint/v1.27.0/golangci-lint

cmd/spire-agent/cli/run/run.go:23: File is not `goimports`-ed (goimports)

	"github.com/sirupsen/logrus"

Makefile:377: recipe for target 'lint-code' failed

[azdagron]

goimports orders the imports alphabetically, which would place "github.com/spiffe/spire/cmd/spire-agent/cli/common" just after "github.com/sirupsen/logrus".

[amartinezfayo]

Thoughts @evan2645 , @amartinezfayo , @APTy , @mcpherrinm? Any concern taking this potentially breaking change? What is your evaluation of the risk?

I'm OK with this change. I see the risk of breaking something really small and I think that it's good to have this fixed before 1.0.
I would make sure that we include this change in the CHANGELOG for the release so anyone updating is aware of the change.

[mcpherrinm]

While I think this is a technically a breaking change, I think it's worth fixing this to a better default now.

[azdagron]

@joewilliams , thanks for putting this together! We (maintainers) huddled and discussed this offline. I think we are ok accepting the small risk. If you wouldn't mind fixing up the sign-off on the commits (and maybe squash and rebase), we're ready to get this merged.

[joewilliams]

@azdagron okay, got things cleaned up.