http://gi.tibco.com/svnweb/gi/log/trunk 2007-12-12T02:24:34 SVN::Web daily 1 1901-01-01T00:00+00:00 http://gi.tibco.com/svnweb/gi/revision?rev=664 weingram 2007-12-12T02:24:34 http://gi.tibco.com/svnweb/gi/revision?rev=663 weingram 2007-12-12T00:32:35 http://gi.tibco.com/svnweb/gi/revision?rev=662 weingram 2007-12-11T02:34:28 http://gi.tibco.com/svnweb/gi/revision?rev=661
* Renamed Registry as "Repository"
* Fixed minor typo in test script
]]> weingram 2007-11-28T02:08:41 http://gi.tibco.com/svnweb/gi/revision?rev=660
* Reorganized samples directory into subdirectories
* Added index page for the samples
* Added simplified store/query sample
* Significantly enhanced original store/query sample to show 2 subscribers and buttons rather than 1 subscriber and links (a much clearer example of the use case)]]> weingram 2007-11-27T22:20:33 http://gi.tibco.com/svnweb/gi/revision?rev=659 weingram 2007-11-27T22:17:11 http://gi.tibco.com/svnweb/gi/revision?rev=658 weingram 2007-11-19T23:05:46 http://gi.tibco.com/svnweb/gi/revision?rev=657
Added Functional Test 4.x, for PageBus store and query functions ("registry").

Tested on IE 6 and 7, FireFox 2, Safari for Windows Beta 3 and Opera 9 for Windows.

Functional tests pass on all platforms.

JSUnit performance test for Safari on Windows has an issue. The non-JSUnit performance tests work fine on all of the browsers. ]]> weingram 2007-11-15T21:42:14 http://gi.tibco.com/svnweb/gi/revision?rev=656 weingram 2007-11-15T19:42:18 http://gi.tibco.com/svnweb/gi/revision?rev=655 weingram 2007-11-15T19:26:50 http://gi.tibco.com/svnweb/gi/revision?rev=654 weingram 2007-11-15T15:54:02 http://gi.tibco.com/svnweb/gi/revision?rev=653 dhwang 2007-10-23T23:29:18 http://gi.tibco.com/svnweb/gi/revision?rev=652
3 comments:

1.publish only stores the first error to occur during the publish; it throws one error, not an array of them. The 1st error is most likely to be a root cause of something that could cause other errors later on in one of the other callbacks. However, if other independent subscriber errors occur, these are silently swallowed.
2. The current spec & test suite do not distinguish this behavior from the prior behavior, which was to immediately abort the publish on the 1st exception without calling subsequent callbacks.
3. It is a DOCUMENTED BEST PRACTICE for production-grade subscriber callbacks NEVER to throw exceptions. They should always catch and handle exceptions themselves. The exception handling here is only going to come into play if developers fail to follow this simple rule! The exception mechanism here offers a little bit of protection for applications when poorly written subscribers are involved. ]]> weingram 2007-10-18T05:51:12 http://gi.tibco.com/svnweb/gi/revision?rev=651 ]]> weingram 2007-10-16T09:32:08 http://gi.tibco.com/svnweb/gi/revision?rev=650 * embedded OpenAjax Hub
* PageBus impl now calls the OAH funcs
* added filter function to subscribe ]]> weingram 2007-10-16T09:31:12 http://gi.tibco.com/svnweb/gi/revision?rev=602 jcostell 2007-09-04T20:26:02 http://gi.tibco.com/svnweb/gi/revision?rev=601 GI-256 (1-8G0TWX) TimePicker.getValue()]]> jcostell 2007-09-04T19:01:47 http://gi.tibco.com/svnweb/gi/revision?rev=600 jcostell 2007-09-04T17:10:23 http://gi.tibco.com/svnweb/gi/revision?rev=599 jcostell 2007-08-30T02:25:03 http://gi.tibco.com/svnweb/gi/revision?rev=598 jcostell 2007-08-22T17:17:46