nil pointer dereference in tls check in v0.4.1

[dangogh]

Describe the bug
Trying to use a minimal tls.Config, I get this crash:

panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0xb74d33]

goroutine 38 [running]:
github.com/segmentio/kafka-go.(*connGroup).connect(0xc0001f3480, 0xed4dc0, 0xc000077620, 0x0, 0x0, 0x0)
	/go/pkg/mod/github.com/segmentio/[email protected]/transport.go:1016 +0xc63
github.com/segmentio/kafka-go.(*connGroup).grabConnOrConnect.func1(0xc0001f3480, 0xed4d40, 0xc0001f3400, 0xc000101e00, 0xc000101da0)
	/go/pkg/mod/github.com/segmentio/[email protected]/transport.go:881 +0x6e
created by github.com/segmentio/kafka-go.(*connGroup).grabConnOrConnect
	/go/pkg/mod/github.com/segmentio/[email protected]/transport.go:880 +0x101

Kafka Version
2.2.1 (MSK on AWS)

To Reproduce
I wrote this test to show the crash:

package kafka

import (
	"context"
	"crypto/tls"
	"testing"
)

func TestMakeItCrash(t *testing.T) {
	cg := connGroup{
		pool: &connPool{
			tls: &tls.Config{},
		},
	}

	cg.connect(context.Background())
}

Expected behavior
Successful connection with kafka when a minimal tls.Config is used.

Additional context
It appears the netAddr variable is never being set in this function. It's declared as net.Addr which is an interface, but never set otherwise, so it would be nil.

[achille-roussel]

Thanks for the bug report! I'm looking into this 👍

[achille-roussel]

Note that v0.4.x is not yet officially stable, we're still ironing out the details. If you need a stable version you can pin your dependency at v0.3.7

[dangogh]

thanks! I'll lock on v0.3.7 for a while, then..

[achille-roussel]

Hey @dangogh, I'm going to close this issue because we are tracking the bug in #487 as well, and #504 should be merged soon to address the problem.

Thanks again for reporting!

[dangogh]

@achille-roussel -- works for me! thanks!

…

On Tue, Sep 29, 2020 at 2:39 AM Achille ***@***.***> wrote: Hey @dangogh <https://github.com/dangogh>, I'm going to close this issue because we are tracking the bug in #487 <#487> as well, and #504 <#504> should be merged soon to address the problem. Thanks again for reporting! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#477 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAVXHUFEVISR4QSKMS36YDSIGMMPANCNFSM4P2OQTYA> .

[dangogh]

@achille-roussel -- You mentioned 0.4.x is not yet stable. Is there a rough timetable for official release?

dependabot keeps bugging me to upgrade each time there's a new tag and just want to know when to pull the trigger...

thanks!