OJI Agenda 06 March 2001

- Action items for next week:
- Action items for this meeting

  Ed to give wf.tgz to terry

  Stanley to post findings about bug 68759

  Stanley to give thoughts on the impact of having one nsILiveconnect
  per plugin instance rather than having one nsILiveconnect instance per
  application.

  Ed to visit Joe next Thursday to ascertain where he is on
  understanding liveconnect.

  Ed to start plugin versioning taskforce

  Xiaobin to give Ed list of bugs relating to installation of java.

  Raju to test 1.3.1 java plugin with Netscape 6.01, and 6.0A

  Jeff: what is the bug for RSA source comparison?  Stanley says it
  won't be in JDK 1.4.

  Joe to write round-trip testcase.

  Joe to report on status of 70588 nsILiveConnect cardinality

  Raju to find out about SQA's liveconnect test base.

- Action items from previous meetings:

- Agenda:

Can we have a definitive understanding of what is broken in liveconnect?

Can we have a status statement on installation problems?

Discussed workarounds for RSA source comparison issue:

  Jeff: what is the bug for RSA source comparison?  Stanley says it
  won't be in JDK 1.4.

  The code is in lcglue.cpp: jsj_enter_js_impl().  Codebase comparison.
  The problem is that mozilla maintains the certificate one way, but the
  java plugin maintains the certificate as a java array.  Ed suggested
  possibly writing some browser side java code that would turn the java
  array back into the format expected by netscape.  Stanley pointed out
  that when the java side generates the java certificate array, some
  information is lost, such that the original certificate can't be
  re-constituted.  Jeff suggested perhaps we could team with netscape
  people to get to some middle ground.

  The upshot of this bug is that this scenario will not work:

  signed applet communicates with signed javascript.