code review 4437087: net: add network interface identification API

net: add network interface identification API

This CL introduces new API into package net to identify the network
interface.  A functionality of new API is very similar to RFC3493 -
"Interface Identification".

[mikio, 2011-05-27 02:38:58]

Hello golang-dev@googlegroups.com (cc: golang-dev@googlegroups.com),

I'd like you to review this change to
https://go.googlecode.com/hg/

[r, 2011-05-27 05:07:38]

a couple of small points in the portable code.

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:19: if hwa == nil {
if len(hwa) == 0 {

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:25: if i < len(hwa)-1 {
drop the test, always add : and then
return s[:len(s)-1]

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:76: return index, err
for ... {
  if name == ifi.Name {
    return ifi.Index, nil
  }
}
return 0, os.ENOENT

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:96: return name, err
ditto

[mikio, 2011-05-27 05:59:59]

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:19: if hwa == nil {
On 2011/05/27 05:07:38, r wrote:
> if len(hwa) == 0 {

Done.

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:25: if i < len(hwa)-1 {
On 2011/05/27 05:07:38, r wrote:
> drop the test, always add : and then
> return s[:len(s)-1]

Done.

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:76: return index, err
On 2011/05/27 05:07:38, r wrote:
> for ... {
>   if name == ifi.Name {
>     return ifi.Index, nil
>   }
> }
> return 0, os.ENOENT

Done.

http://codereview.appspot.com/4437087/diff/70036/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:96: return name, err
On 2011/05/27 05:07:38, r wrote:
> ditto

Done.

[mikio, 2011-05-27 06:00:16]

Hello golang-dev@googlegroups.com, r@golang.org (cc:
golang-dev@googlegroups.com),

Please take another look.

[gri, 2011-05-27 17:24:16]

FYI.

http://codereview.appspot.com/4437087/diff/83001/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/83001/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:23: for i := 0; i < len(hwa); i++ {
for _, b := range hwa { ...

Also it would probably better to use a bytes.Buffer here (not that it really
matters):

var buf bytes.Buffer
for i, b := range hwa {
   if i > 0 {
      buf.WriteByte(':')
   }
   fmt.Fprintf(&buf, "%02x", b)
}
return buf.String()

http://codereview.appspot.com/4437087/diff/83001/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:54: // InterfaceIndex maps the network interface name
specified in
s/in/by/ ?

http://codereview.appspot.com/4437087/diff/83001/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:72: // InterfaceName maps the network interface index
speficied in
s/speficied/specified/
s/in/by/

[mikio, 2011-05-27 20:39:34]

http://codereview.appspot.com/4437087/diff/83001/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/83001/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:23: for i := 0; i < len(hwa); i++ {
On 2011/05/27 17:24:16, gri wrote:
> for _, b := range hwa { ...
> 
> Also it would probably better to use a bytes.Buffer here (not that it really
> matters):
> 
> var buf bytes.Buffer
> for i, b := range hwa {
>    if i > 0 {
>       buf.WriteByte(':')
>    }
>    fmt.Fprintf(&buf, "%02x", b)
> }
> return buf.String()

better idiomatic things are welcome.
Done.

http://codereview.appspot.com/4437087/diff/83001/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:54: // InterfaceIndex maps the network interface name
specified in
On 2011/05/27 17:24:16, gri wrote:
> s/in/by/ ?

Done.

http://codereview.appspot.com/4437087/diff/83001/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:72: // InterfaceName maps the network interface index
speficied in
On 2011/05/27 17:24:16, gri wrote:
> s/speficied/specified/
> s/in/by/

Done.

[mikio, 2011-05-27 20:40:08]

Hello golang-dev@googlegroups.com, r@golang.org, gri@golang.org (cc:
golang-dev@googlegroups.com),

Please take another look.

[mikio, 2011-05-27 20:43:44]

Hello golang-dev@googlegroups.com, r@golang.org, gri@golang.org (cc:
golang-dev@googlegroups.com),

Please take another look.

[gri, 2011-05-27 21:47:12]

FYI.

Leaving this to somebody who actually understands network interfaces.
- gri

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.go
File src/pkg/net/interface_stub.go (right):

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:14: // network interfaces.  Otheriwse
interfaceTable returns a mapping
s/Otheriwse interfaceTable/Otherwise it/

Apply same corrections in all files.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:15: // of specific interface.
s/of/of a/

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:21: // on all network interfaces.  Otherwise
interfaceAddrTable
s/on/of/ or /for/ ? not sure what this should be, or if "on" is correct

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:21: // on all network interfaces.  Otherwise
interfaceAddrTable
s/Otherwise interfaceAddrTable/Otherwise it/

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:22: // returns addresses on specific interface.
s/on/on a/ (or of a, for a, - same as above)

[bradfitz, 2011-05-27 21:54:44]

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:35: Name         string
e.g. "eth0" ?

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:36: Type         string
what is type? worth enumerating? could use docs in any case.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:45: func EnumerateInterface() ([]Interface, os.Error) {
Enumerate seems like a big word for Go.

I wonder if the other Gophers would prefer just "Interfaces" for this.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:84: if index == ifi.Index {
you already passed index to interfaceTable(index)?  is this check necessary too?

[mikio, 2011-05-29 05:39:48]

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:35: Name         string
On 2011/05/27 21:54:44, bradfitz wrote:
> e.g. "eth0" ?

Done.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:36: Type         string
On 2011/05/27 21:54:44, bradfitz wrote:
> what is type? worth enumerating? could use docs in any case.

Type shows media or link-type. I agree with you, it's not useful 
information because if someone needs more accurate things they
will use MIB. I realized that what I would like to see is Flags, 
which contains state of interface, type of interface, not media/link 
type.

Done.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:45: func EnumerateInterface() ([]Interface, os.Error) {
On 2011/05/27 21:54:44, bradfitz wrote:
> Enumerate seems like a big word for Go.
> 
> I wonder if the other Gophers would prefer just "Interfaces" for this.

Thanks, changed.
Done.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:84: if index == ifi.Index {
On 2011/05/27 21:54:44, bradfitz wrote:
> you already passed index to interfaceTable(index)?  is this check necessary
too?

Removed redundant check.
Done.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.go
File src/pkg/net/interface_stub.go (right):

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:14: // network interfaces.  Otheriwse
interfaceTable returns a mapping
On 2011/05/27 21:47:12, gri wrote:
> s/Otheriwse interfaceTable/Otherwise it/
> 
> Apply same corrections in all files.

Done.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:15: // of specific interface.
On 2011/05/27 21:47:12, gri wrote:
> s/of/of a/

Done.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:21: // on all network interfaces.  Otherwise
interfaceAddrTable
On 2011/05/27 21:47:12, gri wrote:
> s/on/of/ or /for/ ? not sure what this should be, or if "on" is correct

Done.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:21: // on all network interfaces.  Otherwise
interfaceAddrTable
On 2011/05/27 21:47:12, gri wrote:
> s/Otherwise interfaceAddrTable/Otherwise it/

Done.

http://codereview.appspot.com/4437087/diff/56040/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:22: // returns addresses on specific interface.
On 2011/05/27 21:47:12, gri wrote:
> s/on/on a/ (or of a, for a, - same as above)

Done.

[mikio, 2011-05-29 05:40:38]

Hello golang-dev@googlegroups.com, r@golang.org, gri@golang.org,
bradfitz@golang.org (cc: golang-dev@googlegroups.com),

Please take another look.

[mikio, 2011-06-01 10:01:16]

Ping.

On Sun, May 29, 2011 at 2:40 PM,  <mikioh.mikioh@gmail.com> wrote:
> Hello golang-dev@googlegroups.com, r@golang.org, gri@golang.org,
> bradfitz@golang.org (cc: golang-dev@googlegroups.com),
>
> Please take another look.
>
>
> http://codereview.appspot.com/4437087/
>

[rh, 2011-06-02 04:52:52]

What about remove interface_stub?

All OSes get interface_$(GOOS).go in Makefile.

Plan 9 gets: 

func interfaceTable(ifindex int) ([]Interface, os.Error) {
         return nil, os.EPLAN9
}...

Windows gets:

func interfaceTable(ifindex int) ([]Interface, os.Error) {
         return nil, os.NewSyscallError("interfaceTable", syscall.EWINDOWS)
}...

or something like this?

[mikio, 2011-06-02 05:55:37]

On Thu, Jun 2, 2011 at 1:52 PM,  <robert.hencke@gmail.com> wrote:

> What about remove interface_stub?
>
> All OSes get interface_$(GOOS).go in Makefile.

I'd rather not go that way because a) already sendfile_stub.go and
cgo_stub.go exist, b) we must have more two shim files for both
Darwin and FreeBSD.

[albert.strasheim, 2011-06-02 09:42:55]

On 2011/05/27 02:38:58, mikioh wrote:
> Hello mailto:golang-dev@googlegroups.com (cc:
mailto:golang-dev@googlegroups.com),
> 
> I'd like you to review this change to
> https://go.googlecode.com/hg/

It is possible to be in love with a CL?

If so, I am.

Regards

Albert

P.S. Tested this patch on a Linux machine with both Ethernet and IPoIB
interfaces and it works very well.

[rsc, 2011-06-02 15:13:32]

Looks good; main concern is just error handling.
I am a little worried about overloading the word 'interface'
but I don't see a better option.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:15: // HardwareAddr represents the hardware address of
the interface.
// A HardwareAddr represents a physical hardware address.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:18: // String returns the string form of the hardware
address.
delete; we typically don't bother with this one unless
there's something interesting to add.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:19: func (hwa HardwareAddr) String() string {
s/hwa/a/

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:36: Flags        int          // e.g., syscall.IFF_UP,
syscall.IFF_MULTICAST
It would be nice if you didn't have to depend on syscall for the flag bits.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:42: func (ifi *Interface) Addr() ([]IP, os.Error) {
Shouldn't this return []Addr?

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:52: func InterfaceAddrs() ([]IP, os.Error) {
Same?

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:56: // InterfaceIndex maps the network interface name
specified by
I think these should be

func InterfaceByIndex(index int) (*Interface, os.Error)
func InterfaceByName(name string) (*Interface, os.Error)

Then you can find out more than just one field.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:60: return 0, os.EINVAL
The errors need detail.  See other errors in this package.
Applies throughout.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface_bsd.go
File src/pkg/net/interface_bsd.go (right):

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface_bsd.go...
src/pkg/net/interface_bsd.go:90: return "other"
return "media" + itoa(int(media))

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface_stub.go
File src/pkg/net/interface_stub.go (right):

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:9: import (
can drop ( )

[mikio, 2011-06-03 15:01:40]

All done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:15: // HardwareAddr represents the hardware address of
the interface.
On 2011/06/02 15:13:32, rsc wrote:
> // A HardwareAddr represents a physical hardware address.

Done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:18: // String returns the string form of the hardware
address.
On 2011/06/02 15:13:32, rsc wrote:
> delete; we typically don't bother with this one unless
> there's something interesting to add.

Done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:19: func (hwa HardwareAddr) String() string {
On 2011/06/02 15:13:32, rsc wrote:
> s/hwa/a/

Done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:36: Flags        int          // e.g., syscall.IFF_UP,
syscall.IFF_MULTICAST
On 2011/06/02 15:13:32, rsc wrote:
> It would be nice if you didn't have to depend on syscall for the flag bits.

Agreed, added methods to Interface instead.
Done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:42: func (ifi *Interface) Addr() ([]IP, os.Error) {
On 2011/06/02 15:13:32, rsc wrote:
> Shouldn't this return []Addr?

Done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:52: func InterfaceAddrs() ([]IP, os.Error) {
On 2011/06/02 15:13:32, rsc wrote:
> Same?

Done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:56: // InterfaceIndex maps the network interface name
specified by
On 2011/06/02 15:13:32, rsc wrote:
> I think these should be
> 
> func InterfaceByIndex(index int) (*Interface, os.Error)
> func InterfaceByName(name string) (*Interface, os.Error)
> 
> Then you can find out more than just one field.

Makes sense.
Done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:60: return 0, os.EINVAL
On 2011/06/02 15:13:32, rsc wrote:
> The errors need detail.  See other errors in this package.
> Applies throughout.

Done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface_bsd.go
File src/pkg/net/interface_bsd.go (right):

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface_bsd.go...
src/pkg/net/interface_bsd.go:90: return "other"
On 2011/06/02 15:13:32, rsc wrote:
> return "media" + itoa(int(media))

At this point linkType is not needed, removed.
Done.

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface_stub.go
File src/pkg/net/interface_stub.go (right):

http://codereview.appspot.com/4437087/diff/64044/src/pkg/net/interface_stub.g...
src/pkg/net/interface_stub.go:9: import (
On 2011/06/02 15:13:32, rsc wrote:
> can drop ( )

Done.

[mikio, 2011-06-03 15:02:05]

Hello r@golang.org, gri@golang.org, bradfitz@golang.org,
robert.hencke@gmail.com, fullung@gmail.com, rsc@golang.org (cc:
golang-dev@googlegroups.com),

Please take another look.

[rsc, 2011-06-03 16:50:07]

LGTM

Please make the changes below and I'll submit it.

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:57: // SupportsBroadcast returns true if ifi supports a
broadcast
CanBroadcast

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:75: // SupportsMulticast returns true if ifi supports a
multicast
CanMulticast

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:87: return nil, os.NewError("netif: invalid interface")
These should all say "net:" (the package name)

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:92: // Interfaces tries to enumerate network
interfaces.
Interfaces returns a list of the system's network interfaces.

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:97: // InterfaceAddrs tries to enumerate network
interface addresses.
Interfaces returns a list of the system's network interface addresses.

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:102: // InterfaceByIndex returns an interface specified
by index.
s/an/the/

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:117: // InterfaceByName returns an interface specified
by name.
s/an/the/

[mikio, 2011-06-03 18:18:10]

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go
File src/pkg/net/interface.go (right):

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:57: // SupportsBroadcast returns true if ifi supports a
broadcast
On 2011/06/03 16:50:07, rsc wrote:
> CanBroadcast

Done.

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:75: // SupportsMulticast returns true if ifi supports a
multicast
On 2011/06/03 16:50:07, rsc wrote:
> CanMulticast

Done.

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:87: return nil, os.NewError("netif: invalid interface")
On 2011/06/03 16:50:07, rsc wrote:
> These should all say "net:" (the package name)

Done.

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:92: // Interfaces tries to enumerate network
interfaces.
On 2011/06/03 16:50:07, rsc wrote:
> Interfaces returns a list of the system's network interfaces.

Done.

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:97: // InterfaceAddrs tries to enumerate network
interface addresses.
On 2011/06/03 16:50:07, rsc wrote:
> Interfaces returns a list of the system's network interface addresses.

Done.

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:102: // InterfaceByIndex returns an interface specified
by index.
On 2011/06/03 16:50:07, rsc wrote:
> s/an/the/

Done.

http://codereview.appspot.com/4437087/diff/63059/src/pkg/net/interface.go#new...
src/pkg/net/interface.go:117: // InterfaceByName returns an interface specified
by name.
On 2011/06/03 16:50:07, rsc wrote:
> s/an/the/

Done.

[mikio, 2011-06-03 18:18:27]

Hello r@golang.org, gri@golang.org, bradfitz@golang.org,
robert.hencke@gmail.com, fullung@gmail.com, rsc@golang.org (cc:
golang-dev@googlegroups.com),

Please take another look.

[rsc, 2011-06-03 18:35:44]

*** Submitted as http://code.google.com/p/go/source/detail?r=9f4046737460 ***

net: add network interface identification API

This CL introduces new API into package net to identify the network
interface.  A functionality of new API is very similar to RFC3493 -
"Interface Identification".

R=r, gri, bradfitz, robert.hencke, fullung, rsc
CC=golang-dev
http://codereview.appspot.com/4437087

Committer: Russ Cox <rsc@golang.org>

[albert.strasheim, 2011-06-08 15:32:55]

Hello

On 2011/06/03 18:35:44, rsc wrote:
> *** Submitted as http://code.google.com/p/go/source/detail?r=9f4046737460 ***
> net: add network interface identification API

The rawFlags field that was added to Interface near the end of this CL breaks
code like this:

ifaces, err := net.Interfaces()
if err != nil {
	panic(err)
}
log.Printf("%+v", ifaces)
for _, iface := range ifaces {
	log.Printf("%v", iface.IsUp())
}

Maybe Interface needs to be an interface or Interfaces() should return
*Interface?

Regards

Albert

[mikio, 2011-06-08 16:32:59]

Hi,

On Thu, Jun 9, 2011 at 12:32 AM,  <fullung@gmail.com> wrote:

> The rawFlags field that was added to Interface near the end of this CL
> breaks code like this:
(snip)
> Maybe Interface needs to be an interface or Interfaces() should return
> *Interface?

Does this make sense to you?
<http://codereview.appspot.com/4602044/>

[albert.strasheim, 2011-06-08 16:40:21]

Hello

On Wed, Jun 8, 2011 at 6:32 PM, Mikio Hara <mikioh.mikioh@gmail.com> wrote:
> Hi,
> On Thu, Jun 9, 2011 at 12:32 AM,  <fullung@gmail.com> wrote:
>> The rawFlags field that was added to Interface near the end of this CL
>> breaks code like this:
> (snip)
>> Maybe Interface needs to be an interface or Interfaces() should return
>> *Interface?
> Does this make sense to you?
> <http://codereview.appspot.com/4602044/>

LGTM. I'll update my patch for IsMaster and IsSlave if this gets accepted.

Regards

Albert

[kevlar, 2011-06-10 16:01:17]

I see the following issue on two separate Linux amd_64 boxes at tip:

http://code.google.com/p/go/issues/detail?id=1942

On Wed, Jun 8, 2011 at 9:32 AM, Mikio Hara <mikioh.mikioh@gmail.com> wrote:

> Hi,
>
> On Thu, Jun 9, 2011 at 12:32 AM,  <fullung@gmail.com> wrote:
>
> > The rawFlags field that was added to Interface near the end of this CL
> > breaks code like this:
> (snip)
> > Maybe Interface needs to be an interface or Interfaces() should return
> > *Interface?
>
> Does this make sense to you?
> <http://codereview.appspot.com/4602044/>
>

[mikio, 2011-06-10 22:01:32]

On Sat, Jun 11, 2011 at 1:00 AM, Kyle Lemons <kevlar@google.com> wrote:

> I see the following issue on two separate Linux amd_64 boxes at tip:
> http://code.google.com/p/go/issues/detail?id=1942

Thank you for the report and nice %q.
The issue will be fixed by CL 462044.