[lsb-discuss] Java

[Jeff Licquia]

Theodore Tso wrote:
> On Fri, Feb 01, 2008 at 12:02:50PM -0500, Robert Schweikert wrote:
>> For this use case we need to further explore the "plugin certification 
>> model" we have kicked around. Basically the vendors makes the statement 
>> that says "Certified with any LSB compliant JRE". How we will test this 
>> type of app needs to be determined, not sure whether or not Jeff had a 
>> proposal for this.
> 
> Yeah... I'm a little concerned about this.  The devil will be in the
> details.

The proposal is on my task list for post-3.2 stuff.  Here's an informal 
summary:

In LSB 4.0, ELF binaries can be certified with a "plugin addendum", 
whereby the ELF binary is claimed to be a plugin for some other system. 
  Thus, failures to resolve symbols in the ELF binary from within the 
app or from within the LSB would be waived, provided the symbols met 
some qualification.  Loosely speaking, that qualification would be tied 
to some requirement to document the plugin ABI.

The original use case for this proposal was the ever-popular Mozilla 
plugin API, although the idea was to make this usable for other popular 
plugin systems, such as GStreamer or widget theme engines.

As observed, there are some traps to avoid when granting approval to 
plugin ABIs, which is why the proposal hasn't been made just yet.

IMHO, the plugin proposal is somewhat orthogonal to the question of 
certifying Java apps.  Certainly, if we can create a Java runtime that 
is certified to the LSB and that runs the Java app in question, the 
combination of the two should be certifiable as a LSB application today, 
without the plugin proposal.

The value here may be in creating LSB-certified JREs for each of our 
architectures that can also be certified to Sun's program. 
Theoretically, Sun's certification could be translated to Linux 
deployment, though we all know that logic isn't everything in these 
situations.