Add new InternetAddress class with a static lookup function (including IPv6 results).

Add new InternetAddress class with a static lookup function (including IPv6 results).

Socket.connect/ServerSocket.bind now accepts IPv6 host names. If a new ServerSocket is connected to an IPv6 interface, it will (if possible) support incoming IPv4 connections as well.

BUG=

Committed: https://code.google.com/p/dart/source/detail?r=21862

[Anders Johnsen, 2013-04-18 08:18:06]

Very early version.

- No tests
- No mac version
- No win version
- No android version

But I would like some comments on C++ / overall API.

[Søren Gjesse, 2013-04-18 09:07:40]

Initial comments

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h
File runtime/bin/socket.h (right):

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h#newcode44
runtime/bin/socket.h:44: Type type;
Make these mempers private with _ postfix. Then add getters.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h#newcode46
runtime/bin/socket.h:46: uint8_t* data;
can't this be

sockaddr_storage address

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h#newcode47
runtime/bin/socket.h:47: const intptr_t length;
Please add:

 private:
  DISALLOW_COPY_AND_ASSIGN(SocketAddress);

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h#newcode62
runtime/bin/socket.h:62: 
As above.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_linux.cc
File runtime/bin/socket_linux.cc (right):

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_linux.cc#n...
runtime/bin/socket_linux.cc:28: intptr_t
Socket::CreateConnect(SocketAddress::Type type,
I don't think we need the type. sockaddr_in has a family member first.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_linux.cc#n...
runtime/bin/socket_linux.cc:29: uint8_t* host_data,
sockaddr_storage* address or sockaddr_in* address

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_linux.cc#n...
runtime/bin/socket_linux.cc:178: if (c->ai_family == AF_INET || c->ai_family ==
AF_INET6) {
How about moving all this into the SocketAddress constructor?

SocketAddress(struct addrinfo* addrinfo) {
  ...
}

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_patch.dart
File runtime/bin/socket_patch.dart (right):

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_patch.dart...
runtime/bin/socket_patch.dart:30: final List<int> _data;
I don't like having this here, but I guess keeoing it in C using a class which
extends NativeFieldWrapperClass1 will get complicated.

It should be typed Uint8List and called _sockaddr_in.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_patch.dart...
runtime/bin/socket_patch.dart:119: static Future<_NativeSocket> connect(host,
int port) {
Should we add type checks now that checked mode will not give any warnings?

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart
File sdk/lib/io/socket.dart (right):

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart#newcode10
sdk/lib/io/socket.dart:10: static const InternetAddressType IPv6 = const
InternetAddressType._("IPv6");
Change this to an integer with the same values as the C SocketAddress::Type
enum. The you don't need the   ... ? 0 : 1 on the other files, but can use a
value getter.

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart#newcode12
sdk/lib/io/socket.dart:12: const InternetAddressType._(String this.name);
Please ass toString as well.

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart#newcod...
sdk/lib/io/socket.dart:161: * if the host-lookup or connection failed.
Update comment on type of host.

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart#newcod...
sdk/lib/io/socket.dart:250: * if the host-lookup or connection failed.
Ditto.

[Anders Johnsen, 2013-04-18 12:09:45]

Thanks a ton for the comments Soeren, much nicer now! PTAL before I start adding
tests and porting to other platforms.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h
File runtime/bin/socket.h (right):

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h#newcode44
runtime/bin/socket.h:44: Type type;
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> Make these mempers private with _ postfix. Then add getters.

Done.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h#newcode46
runtime/bin/socket.h:46: uint8_t* data;
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> can't this be
> 
> sockaddr_storage address

Done.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h#newcode47
runtime/bin/socket.h:47: const intptr_t length;
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> Please add:
> 
>  private:
>   DISALLOW_COPY_AND_ASSIGN(SocketAddress);

Done.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket.h#newcode62
runtime/bin/socket.h:62: 
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> As above.

Done.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_linux.cc
File runtime/bin/socket_linux.cc (right):

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_linux.cc#n...
runtime/bin/socket_linux.cc:28: intptr_t
Socket::CreateConnect(SocketAddress::Type type,
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> I don't think we need the type. sockaddr_in has a family member first.

Done.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_linux.cc#n...
runtime/bin/socket_linux.cc:29: uint8_t* host_data,
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> sockaddr_storage* address or sockaddr_in* address

Done.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_linux.cc#n...
runtime/bin/socket_linux.cc:178: if (c->ai_family == AF_INET || c->ai_family ==
AF_INET6) {
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> How about moving all this into the SocketAddress constructor?
> 
> SocketAddress(struct addrinfo* addrinfo) {
>   ...
> }

Done.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_patch.dart
File runtime/bin/socket_patch.dart (right):

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_patch.dart...
runtime/bin/socket_patch.dart:30: final List<int> _data;
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> I don't like having this here, but I guess keeoing it in C using a class which
> extends NativeFieldWrapperClass1 will get complicated.
> 
> It should be typed Uint8List and called _sockaddr_in.

Done.

https://codereview.chromium.org/14083007/diff/1/runtime/bin/socket_patch.dart...
runtime/bin/socket_patch.dart:119: static Future<_NativeSocket> connect(host,
int port) {
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> Should we add type checks now that checked mode will not give any warnings?

The runtime-call will actually do the check for us.

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart
File sdk/lib/io/socket.dart (right):

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart#newcode10
sdk/lib/io/socket.dart:10: static const InternetAddressType IPv6 = const
InternetAddressType._("IPv6");
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> Change this to an integer with the same values as the C SocketAddress::Type
> enum. The you don't need the   ... ? 0 : 1 on the other files, but can use a
> value getter.

Done.

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart#newcode12
sdk/lib/io/socket.dart:12: const InternetAddressType._(String this.name);
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> Please ass toString as well.

Done.

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart#newcod...
sdk/lib/io/socket.dart:161: * if the host-lookup or connection failed.
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> Update comment on type of host.

Done.

https://codereview.chromium.org/14083007/diff/1/sdk/lib/io/socket.dart#newcod...
sdk/lib/io/socket.dart:250: * if the host-lookup or connection failed.
On 2013/04/18 09:07:40, Søren Gjesse wrote:
> Ditto.

Done.

[Søren Gjesse, 2013-04-18 12:59:24]

The code looks good now.

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket.h
File runtime/bin/socket.h (right):

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket.h#newco...
runtime/bin/socket.h:46: const char* as_string_;
How about static buffer for as_string as well?

char as_string[INET6_ADDRSTRLEN];

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_linux.cc
File runtime/bin/socket_linux.cc (right):

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_linux.c...
runtime/bin/socket_linux.cc:55: reinterpret_cast<struct
sockaddr_in*>(&addr)->sin_port = htons(port);
We better switch on family here

if (addr.ss_family == AF_INET) {
  reinterpret_cast<struct sockaddr_in*>(&addr)->sin_port = htons(port);
} else {
  reinterpret_cast<struct sockaddr_in6*>(&addr)->sin6_port = 
htons(port);
}

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_linux.c...
runtime/bin/socket_linux.cc:170: hints.ai_protocol = 0;
Don't you still want only IPPROTO_TCP?

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_patch.dart
File runtime/bin/socket_patch.dart (right):

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_patch.d...
runtime/bin/socket_patch.dart:104: .then((response) {
You better add a comment here that this is checking for OS error. Consider
changing the message format to have an integer as the first element in the array
always indicating success or failure. common.dart have the constants used in
other places.

in file_impl.dart all RPC calls looks like this:

    List request = new List(...);
    // Fill request.
    return _fileService.call(request).then((response) {
      if (_isErrorResponse(response)) {
        throw _exceptionFromResponse(response, ...);
      }
      return ...
    });

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_patch.d...
runtime/bin/socket_patch.dart:108: return [];
Throw an OSError here.

[Anders Johnsen, 2013-04-19 09:45:36]

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket.h
File runtime/bin/socket.h (right):

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket.h#newco...
runtime/bin/socket.h:46: const char* as_string_;
On 2013/04/18 12:59:24, Søren Gjesse wrote:
> How about static buffer for as_string as well?
> 
> char as_string[INET6_ADDRSTRLEN];

Done.

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_linux.cc
File runtime/bin/socket_linux.cc (right):

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_linux.c...
runtime/bin/socket_linux.cc:55: reinterpret_cast<struct
sockaddr_in*>(&addr)->sin_port = htons(port);
On 2013/04/18 12:59:24, Søren Gjesse wrote:
> We better switch on family here
> 
> if (addr.ss_family == AF_INET) {
>   reinterpret_cast<struct sockaddr_in*>(&addr)->sin_port = htons(port);
> } else {
>   reinterpret_cast<struct sockaddr_in6*>(&addr)->sin6_port = 
> htons(port);
> }

Done.

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_linux.c...
runtime/bin/socket_linux.cc:170: hints.ai_protocol = 0;
On 2013/04/18 12:59:24, Søren Gjesse wrote:
> Don't you still want only IPPROTO_TCP?

Done.

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_patch.dart
File runtime/bin/socket_patch.dart (right):

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_patch.d...
runtime/bin/socket_patch.dart:104: .then((response) {
On 2013/04/18 12:59:24, Søren Gjesse wrote:
> You better add a comment here that this is checking for OS error. Consider
> changing the message format to have an integer as the first element in the
array
> always indicating success or failure. common.dart have the constants used in
> other places.
> 
> in file_impl.dart all RPC calls looks like this:
> 
>     List request = new List(...);
>     // Fill request.
>     return _fileService.call(request).then((response) {
>       if (_isErrorResponse(response)) {
>         throw _exceptionFromResponse(response, ...);
>       }
>       return ...
>     });

Done.

https://codereview.chromium.org/14083007/diff/8001/runtime/bin/socket_patch.d...
runtime/bin/socket_patch.dart:108: return [];
On 2013/04/18 12:59:24, Søren Gjesse wrote:
> Throw an OSError here.

Done.

[Søren Gjesse, 2013-04-19 10:36:06]

The API and Linux code looks good now.

https://codereview.chromium.org/14083007/diff/15001/runtime/bin/socket.h
File runtime/bin/socket.h (right):

https://codereview.chromium.org/14083007/diff/15001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:26: #define SOCKADDR_STORAGE_LENGTH(sa) \
Maybe make this a static method on SocketAddress instead of a macro.

https://codereview.chromium.org/14083007/diff/15001/runtime/bin/socket_linux.cc
File runtime/bin/socket_linux.cc (right):

https://codereview.chromium.org/14083007/diff/15001/runtime/bin/socket_linux....
runtime/bin/socket_linux.cc:28: SocketAddress::SocketAddress(struct addrinfo*
addrinfo) {
Add the assert

ASSERT(INET6_ADDRSTRLEN >= INET_ADDRSTRLEN);

[Anders Johnsen, 2013-04-22 14:05:15]

PTAL

https://codereview.chromium.org/14083007/diff/15001/runtime/bin/socket.h
File runtime/bin/socket.h (right):

https://codereview.chromium.org/14083007/diff/15001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:26: #define SOCKADDR_STORAGE_LENGTH(sa) \
On 2013/04/19 10:36:06, Søren Gjesse wrote:
> Maybe make this a static method on SocketAddress instead of a macro.

Done.

https://codereview.chromium.org/14083007/diff/15001/runtime/bin/socket_linux.cc
File runtime/bin/socket_linux.cc (right):

https://codereview.chromium.org/14083007/diff/15001/runtime/bin/socket_linux....
runtime/bin/socket_linux.cc:28: SocketAddress::SocketAddress(struct addrinfo*
addrinfo) {
On 2013/04/19 10:36:06, Søren Gjesse wrote:
> Add the assert
> 
> ASSERT(INET6_ADDRSTRLEN >= INET_ADDRSTRLEN);

Done.

[Søren Gjesse, 2013-04-22 15:02:08]

LGTM with Mac and Android pending.

The support for SocketOption IPV6_V6ONLY can be postpone to a separate CL.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.cc
File runtime/bin/socket.cc (right):

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.cc#new...
runtime/bin/socket.cc:27: Dart_Handle host_obj = Dart_GetNativeArgument(args,
1);
Maybe add a static  helper function:

GetSockAddr(Dart_Handle obj, struct scokaddr_storage* addr);

code is duplicated below.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.cc#new...
runtime/bin/socket.cc:265: intptr_t port = 0;
Add ASSERT(INET6_ADDRSTRLEN >= INET_ADDRSTRLEN)

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h
File runtime/bin/socket.h (right):

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:27: public:
Add an enum for type which matched the enum in Dart.

enum {
  TYPE_ANY = -1,
  TYPE_IPV4,
  TYPE_IPV6,
}

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:35: const char* AsString() const { return as_string_; }
This accessor should be called as_string (see
http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml?showone=Functi...).

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:36: const sockaddr_storage& GetAddr() const { return addr_;
}
Ditto, should be called addr.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:42: static int16_t FromType(int type) {
ASSERT type is -1, 0 or 1.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:68: intptr_t GetCount() const { return count_; }
GetCount -> count

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket_linux.cc
File runtime/bin/socket_linux.cc (right):

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket_linux....
runtime/bin/socket_linux.cc:219: optval = 0;
Maybe we should consider a way to keep IPV6_ONLY set (optional argument?).

We should probably add it to SocketOption and support it in Socket.setOption.

We should document that it is on by default.

https://codereview.chromium.org/14083007/diff/25001/sdk/lib/io/secure_socket....
File sdk/lib/io/secure_socket.dart (right):

https://codereview.chromium.org/14083007/diff/25001/sdk/lib/io/secure_socket....
sdk/lib/io/secure_socket.dart:300: bool onBadCertificate(X509Certificate
certificate)}) {
Move this new Future to an else block?

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
File tests/standalone/io/https_client_certificate_test.dart (right):

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
tests/standalone/io/https_client_certificate_test.dart:11: const SERVER_ADDRESS
= "localhost";
Get rid of one of the constants.

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
File tests/standalone/io/https_client_exception_test.dart (right):

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
tests/standalone/io/https_client_exception_test.dart:23: testBadHostName();
Why is this removed?

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
File tests/standalone/io/https_server_test.dart (right):

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
tests/standalone/io/https_server_test.dart:11: const SERVER_ADDRESS =
"localhost";
Remove one of the constants.

In a number of other files as well.

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/socke...
File tests/standalone/io/socket_ipv6_test.dart (right):

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/socke...
tests/standalone/io/socket_ipv6_test.dart:71: if (list.length < 0) throw "no
addresse";
addresse -> address

[Anders Johnsen, 2013-04-23 08:03:57]

PTAL

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.cc
File runtime/bin/socket.cc (right):

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.cc#new...
runtime/bin/socket.cc:27: Dart_Handle host_obj = Dart_GetNativeArgument(args,
1);
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> Maybe add a static  helper function:
> 
> GetSockAddr(Dart_Handle obj, struct scokaddr_storage* addr);
> 
> code is duplicated below.

Done.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.cc#new...
runtime/bin/socket.cc:265: intptr_t port = 0;
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> Add ASSERT(INET6_ADDRSTRLEN >= INET_ADDRSTRLEN)

Done.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h
File runtime/bin/socket.h (right):

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:35: const char* AsString() const { return as_string_; }
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> This accessor should be called as_string (see
>
http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml?showone=Functi...).

Done.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:36: const sockaddr_storage& GetAddr() const { return addr_;
}
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> Ditto, should be called addr.

Done.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:42: static int16_t FromType(int type) {
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> ASSERT type is -1, 0 or 1.

Done.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket.h#newc...
runtime/bin/socket.h:68: intptr_t GetCount() const { return count_; }
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> GetCount -> count

Done.

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket_linux.cc
File runtime/bin/socket_linux.cc (right):

https://codereview.chromium.org/14083007/diff/25001/runtime/bin/socket_linux....
runtime/bin/socket_linux.cc:219: optval = 0;
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> Maybe we should consider a way to keep IPV6_ONLY set (optional argument?).
> 
> We should probably add it to SocketOption and support it in Socket.setOption.
> 
> We should document that it is on by default.

Yeah, good idea. Will do a follow-up on this. I think on by default is a good
choice.

https://codereview.chromium.org/14083007/diff/25001/sdk/lib/io/secure_socket....
File sdk/lib/io/secure_socket.dart (right):

https://codereview.chromium.org/14083007/diff/25001/sdk/lib/io/secure_socket....
sdk/lib/io/secure_socket.dart:300: bool onBadCertificate(X509Certificate
certificate)}) {
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> Move this new Future to an else block?

Done.

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
File tests/standalone/io/https_client_certificate_test.dart (right):

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
tests/standalone/io/https_client_certificate_test.dart:11: const SERVER_ADDRESS
= "localhost";
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> Get rid of one of the constants.

Done.

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
File tests/standalone/io/https_client_exception_test.dart (right):

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
tests/standalone/io/https_client_exception_test.dart:23: testBadHostName();
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> Why is this removed?

It no longer fails. Any ideas why that could be?

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
File tests/standalone/io/https_server_test.dart (right):

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/https...
tests/standalone/io/https_server_test.dart:11: const SERVER_ADDRESS =
"localhost";
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> Remove one of the constants.
> 
> In a number of other files as well.

Done.

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/socke...
File tests/standalone/io/socket_ipv6_test.dart (right):

https://codereview.chromium.org/14083007/diff/25001/tests/standalone/io/socke...
tests/standalone/io/socket_ipv6_test.dart:71: if (list.length < 0) throw "no
addresse";
On 2013/04/22 15:02:08, Søren Gjesse wrote:
> addresse -> address

Done.

[Søren Gjesse, 2013-04-23 09:35:29]

LGTM!

[Anders Johnsen, 2013-04-23 09:36:44]

Committed patchset #11 manually as r21862 (presubmit successful).