2008-07-13Merged 1.0 final into 1.1 branch.
Jens Alfke <jens@mooseyard.com> [Sun, 13 Jul 2008 10:59:06 -0700] rev 23
Merged 1.0 final into 1.1 branch.

2008-07-13* Fixed: Responses still pending when a connection closed were not calling their onComplete targets.
Jens Alfke <jens@mooseyard.com> [Sun, 13 Jul 2008 10:42:50 -0700] rev 22
* Fixed: Responses still pending when a connection closed were not calling their onComplete targets.
* Fixed: BLIPTestClient target failed to build because it didn't link against zlib.
* If TCPListener.bonjourServiceName is changed while the listener is open, it now re-publishes the service with the new name.
* Added a TCPListener.bonjourService property.
* Added a BLIPMessage.representedObject property.
* Fixed a memory leak.

2008-07-13Added tag 1.0 for changeset 02224e981209
Jens Alfke <jens@mooseyard.com> [Sun, 13 Jul 2008 10:53:46 -0700] rev 21
Added tag 1.0 for changeset 02224e981209

2008-07-13* Fixed link error in BLIPClient target. 1.0
Jens Alfke <jens@mooseyard.com> [Sun, 13 Jul 2008 10:52:48 -0700] rev 20
* Fixed link error in BLIPClient target.
* Fixed *Ph*n* #ifdef issue (thanks, Jimmy)

2008-06-23Implemented BLIP 1.1 (explicit 'bye' message)
Jens Alfke <jens@mooseyard.com> [Mon, 23 Jun 2008 14:02:31 -0700] rev 19
Implemented BLIP 1.1 (explicit 'bye' message)

2008-06-19Implemented new close protocol with 'bye' meta-message.
Jens Alfke <jens@mooseyard.com> [Thu, 19 Jun 2008 16:22:05 -0700] rev 18
Implemented new close protocol with 'bye' meta-message.

2008-06-19Work around crash-on-close by temporarily retaining the connection object.
Jens Alfke <jens@mooseyard.com> [Thu, 19 Jun 2008 10:22:19 -0700] rev 17
Work around crash-on-close by temporarily retaining the connection object.

2008-06-11* Added a timeout property to TCPConnection. Set it before calling -open, if you want a shorter timeout than the default.
Jens Alfke <jens@mooseyard.com> [Wed, 11 Jun 2008 14:58:38 -0700] rev 16
* Added a timeout property to TCPConnection. Set it before calling -open, if you want a shorter timeout than the default.
* Made the utility function BLIPMakeError public.

2008-06-10Fixed: The -connection:failedToOpen: delegate method wasn't being called.
Jens Alfke <jens@mooseyard.com> [Tue, 10 Jun 2008 16:14:20 -0700] rev 15
Fixed: The -connection:failedToOpen: delegate method wasn't being called.

2008-06-04Python: Optimized frame sending somewhat (frame buffers are generated on the fly as the socket has room.)
Jens Alfke <jens@mooseyard.com> [Wed, 04 Jun 2008 21:09:45 -0700] rev 14
Python: Optimized frame sending somewhat (frame buffers are generated on the fly as the socket has room.)

2008-06-04Python implementation much improved. Can send requests now. Fully interoperable with Obj-C implementation's test cases.
Jens Alfke <jens@mooseyard.com> [Wed, 04 Jun 2008 17:11:20 -0700] rev 13
Python implementation much improved. Can send requests now. Fully interoperable with Obj-C implementation's test cases.

2008-06-03BLIP.py working for listener side (it talks to the Obj-C BLIPConnectionTester.)
Jens Alfke <jens@mooseyard.com> [Tue, 03 Jun 2008 22:24:21 -0700] rev 12
BLIP.py working for listener side (it talks to the Obj-C BLIPConnectionTester.)

2008-06-03* Initial checkin of BLIP.py. (Receiving seems to work.)
Jens Alfke <jens@mooseyard.com> [Tue, 03 Jun 2008 16:56:33 -0700] rev 11
* Initial checkin of BLIP.py. (Receiving seems to work.)
* FIXED: Abbreviation list in BLIPProperties was messed up.
* Renamed some instance variables to use 'request' instead of 'query'.
* Test client doesn't throw an assertion-failure now when the number of unresponded requests exceeds 100.

2008-06-01Copied the necessary Google Toolbox source files into the MYUtilities project, so people don't have to download a separate library.
Jens Alfke <jens@mooseyard.com> [Sun, 01 Jun 2008 14:04:22 -0700] rev 10
Copied the necessary Google Toolbox source files into the MYUtilities project, so people don't have to download a separate library.

2008-05-30Fixed a serious bug - a race condition where a data buffer in the writer's queue could be dealloced (not the NSData, but its bytes themselves) before the writer sent it, resulting in an EFAULT error.
Jens Alfke <jens@mooseyard.com> [Fri, 30 May 2008 13:54:38 -0700] rev 9
Fixed a serious bug - a race condition where a data buffer in the writer's queue could be dealloced (not the NSData, but its bytes themselves) before the writer sent it, resulting in an EFAULT error.