Socket.IO 0.6.3 released

瀏覽次數:64 次
跳到第一則未讀訊息

Guillermo Rauch

未讀,
2010年12月24日 凌晨12:07:142010/12/24
收件者:sock...@googlegroups.com
Download:

Changelog:

Server:

0.6.3 / 2010-12-23 
==================

  * Changed polling default duration to 50 seconds
  * might > will Adjusted to 85 column limit
  * Support for resources that include slashes. Thanks @schamane
  * Lazy loading of transports. Thanks @technoweenie Fixed README transports list
  * OpenSSL clarifications (thanks @bmnds)
  * Support for HAProxy load balancing (thanks Brian McKelvey) Backported Parser from 0.7
  * Fixed HTTP API in example (was outdated). Thanks deedubs
  * 0.3 compatibility (thanks Arnout)
  * `client.broadcast` now 300% faster Cleaned up chat example 
  * fixed bad pluralization.
  * cleaned up grammar, missing punctuation, etc.
  * Restored global `netserver` for flashsocket Now supporting `flashPolicyServer` option (thanks Arnout) Tests passing with and without sudo/root user Fixed noDelay/timeout/utf-8 for draft 76 (accidental typo)
  * Close the netServer when the main http server closes, this way the event loop does not keep running. NOTE: this is patch for node 0.2.X, this is not required for node 0.3.X
  * Fallback to try{}catch handling for node < 0.2.4 , node 0.3.X seems to capture the errors correctly using the error event.
  * Added the flash policy server, it's enabled by default but can be turned off if needed. Socket.io will automatically fallback to serving the policy file inline if server is disabled or unable to start up.
  * Make sure to only write to open transports (thanks JohnDav)
  * _open is still false, so destroy the connection immediately upon websocket error
  * Make sure .connection is not null on 'end'
  * Proper fix for invalid websocket key

Client:

0.6.1 / 2010-12-13
==================

  * Corrected position of WEB_SOCKET_SWF_LOCATION check
  * Wrapped XHR `abort()` in a try-catch clause
  * Better readyState read
  * send(null) for FF2/3
  * Removed XHR onload handler for faster disconnection detection
  * Fixed connect timeout being fired after an explicit disconnect
  * Renaming fire to emit and making a copy of events first 
  * Fire 'error' event when socket.onerror fires

In addition:
  • For the repository, `master` is now always the latest stable tag, and `develop` is the experimental/future version. 
  • `master` will continue to be the 0.6.x maintenance releases, while develop will contain the upcoming 0.7.
A big thanks and merry christmas to all these amazing contributors (in no specific order):
  • Arnout Kazemier (@3rd-Eden)
  • @schamane
  • @JohnDav
  • Saikat Chakrabati (@saikatc)
  • Rick Olson (@technoweenie)
  • Brian McKelvey
  • Dan Williams (@deedubs)
  • @bmnds
  • @kschzt

Guillermo Rauch
http://devthought.com

Arnout Kazemier

未讀,
2010年12月24日 凌晨4:15:182010/12/24
收件者:sock...@googlegroups.com
Glad you branched the experimental features to a develop branch. And merry christmas :)!

Chris Grindstaff

未讀,
2010年12月24日 上午9:50:322010/12/24
收件者:sock...@googlegroups.com

> *Changelog:*

Looks like a nice set of changes - anyone know if these work with Tad's
Java server? It was my understanding that the plan was to integrate Tad's
client changes back into mainline but I don't think that happened for this
release.

Looking at the diff I don't think it will work but I haven't tried yet.

Thanks,
Chris
--
Chris Grindstaff | http://gstaff.org

Guillermo Rauch

未讀,
2010年12月24日 下午2:02:112010/12/24
收件者:sock...@googlegroups.com
The encoding mechanism in 0.7 is changing to something similar to what Tad came up with, so we'll probably be able to merge the projects at that point.

--
Guillermo Rauch
http://devthought.com
回覆所有人
回覆作者
轉寄
0 則新訊息