W3C home > Mailing lists > Public > public-sysapps@w3.org > June 2013

[sysapps/raw socket api]: Status on outstanding issues and proposal to close a number of issues.

From: Nilsson, Claes1 <Claes1.Nilsson@sonymobile.com>
Date: Thu, 27 Jun 2013 17:47:37 +0200
To: "public-sysapps@w3.org" <public-sysapps@w3.org>
Message-ID: <6DFA1B20D858A14488A66D6EEDF26AA3017DBD1D71E2@seldmbx03.corpusers.net>

The latest version of the Raw Socket API is here: http://www.w3.org/2012/sysapps/raw-sockets/

Open issues are here: https://github.com/sysapps/raw-sockets/issues?state=open

To get an overview I below list the issues and my proposed resolutions. Please note that I propose to close a number of issues if there are no objections.

Support for secure sockets, https://github.com/sysapps/raw-sockets/issues/10:
This is probably the most major remaining task on this API. I am working on this and seeking support from internal security experts. 4D has also promised to help on secure sockets.

Binding to local connection/interface address needed for web applications?, https://github.com/sysapps/raw-sockets/issues/11:
I have stated some use cases that motivate this feature.
If there are no objections I will close the issue.

Joining multicast group for UDP sockets done through separate method or through constructor ?, https://github.com/sysapps/raw-sockets/issues/12:
I suggest to keep the current solution with separate "Join/LeaveMulticastGroup methods".
If there are no objections I will close the issue.

Address reuse for TCPSocket and TCPServerSocket, https://github.com/sysapps/raw-sockets/issues/13:
If there are no objections I will close the issue.

Use Cases and Requirements, https://github.com/sysapps/raw-sockets/issues/14:
I will add this.

Does an application have to join a multicast group even if it is only sending multicast messages, https://github.com/sysapps/raw-sockets/issues/16:
The current wording is based on input from 4D. Should we keep this?

address + port vs uri, https://github.com/sysapps/raw-sockets/issues/17:
I think we need more discussion on this proposal. Personally I need to get a better view on advantages/disadvantages of using URI addresses instead of explicit IP-address and port arguments.

Error handling to be revised, https://github.com/sysapps/raw-sockets/issues/18:
I have made a number of changes that address this issue. See https://github.com/sysapps/raw-sockets/pull/23. Can I close the issue?

Introduce a permission in section 4, based on Runtime and Security spec, https://github.com/sysapps/raw-sockets/issues/21:
If there are no objections I will add the wording I propose to section 4.

async methods should return Future, https://github.com/sysapps/raw-sockets/issues/22:
"Future" is renamed to "promise". Not sure on the approach for the Raw Socket API. Please see the comment thread in the issue.

Expensive to synchronously set TCPSocket.localAddress when using default local address, https://github.com/sysapps/raw-sockets/issues/24:
Any feedback on the two options I propose to address Jonas issue?

Example 1 could be improved, https://github.com/sysapps/raw-sockets/issues/25
Example 2 needs a small cleanup, https://github.com/sysapps/raw-sockets/issues/26:
Expecting input from Marcos

Join/leave group, https://github.com/sysapps/raw-sockets/issues/27:
Personally I would prefer to keep "multicast" in the method names but Marcos wants to drop the word "multicast" and suggest to make the methods extensible. What do others think?

UDPSocket constructor lacks error handling, https://github.com/sysapps/raw-sockets/issues/28:
Is the description ok or does something have to be added? Can I close the issue?

Constructing a UDPSocket must open it, https://github.com/sysapps/raw-sockets/issues/29:
Is the current wording ok? Can I close the issue?

Step 2 in send steps is wrong, https://github.com/sysapps/raw-sockets/issues/30:
I'll remove step 2 if no one objects.

Send() address and port should be args, https://github.com/sysapps/raw-sockets/issues/31:
Closing socket needs to fire a close event. https://github.com/sysapps/raw-sockets/issues/32
I think that the specification is correct here. Marcos please comment. Can I close these issues?

Could send() be a Future?, https://github.com/sysapps/raw-sockets/issues/33
See issue 22.

I am happy for feedback!


Claes Nilsson M.Sc.E.E
Master Engineer - Web Research
Advanced Application Labs

Sony Mobile Communications
Tel: +46 705 56 68 78


(image/jpeg attachment: image001.jpg)

Received on Thursday, 27 June 2013 15:48:08 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:36:13 UTC