Note: This was an informal “unmeeting” this week. Ivar and Kevin attempted to cancel the meeting last minute due the lack of an agenda. But, several people showed up anyway and had a fruitful discussion. Bill Shannon took notes (thank you, Bill!).
Bill Shannon (Oracle)
Emily Jiang (IBM)
David Blevins (Tomitribe)
Vano Beridze (???)
Others were also on the call, but no attendance was recorded or requested…
Despite the meeting being canceled, some people showed up anyway and we have some useful conversations. Don’t consider these minutes, they’re just my recollections of what was talked about, and these are certainly not official minutes. Corrections are encouraged.
Emily Jiang started by asking about the TCKs and lamenting how there was no documentation about the structure of the TCKs and thus it was difficult to determine how to extract pieces of the TCK repo for one particular API in order to create a standalone TCK for that API.
I lamented that we had internal documentation on some of this that we’ve never found time to release. Perhaps some of it would be useful, perhaps not, but I’ll go back and see if there’s any hope of springing it loose of Oracle.
I reminded Emily to talk to Andy Guibert, who has figured out some parts of this.
We talked to David Blevins about EJB and what his plans are for EJB. David was very clear that, given his resource and product requirements, he was not going to be able to have anyone help on the EJB implementation in GlassFish. So, it’s going to be up to the GlassFish project team to do this work.
We talked about the EJB interop tests. David believes there are two kinds of EJB interoperability tests - tests that use only the EJB APIs to test interoperating with remote EJBs, presumably from the same vendor, and tests that use the CORBA APIs to test EJB interoperability using the IIOP, CSIv2, etc. protocols. These later tests should be removed from the TCK.
Note, however, that the only remote EJB protocol supported by GlassFish is the CORBA protocol, so these tests that would be removed from the platform TCK would be very useful as product tests for GlassFish, and we should be careful not to lose them or lose that ability to use them for GlassFish. They would also be useful for any other products that currently provide this level of interoperability and are NOT planning to remove it for Jakarta EE 9 even though it is no longer part of the spec (not even optional).
Vano Beridze lamented that the Jakarta EE RI/CI, Eclipse GlassFish, was clearly not a production level enterprise server, so was some other product going to step in to fill that gap. We explained how there is no longer any RI, how it’s a market competition to determine who is best. We reminded Vano that Payara Server is very close to GlassFish, so contributing to GlassFish may see results in Payara server soon, and for that matter many of the components are shared with other products so it depends on what component you’re contributing to.
Vano was interested in contributing to the Face products. We gave him much of the same “how do I help” advice that has been given on mailing lists several times. Maybe someone (Reza?) could write up a Quick Start page for “how do I get started helping?”
We filled the whole hour with these discussions, and could’ve gone on longer. A few other people joined and left the call and I didn’t try to keep track of them, sorry.
Now I have some advice of my own (not discussed during the meeting)…
If you really want to cancel the meeting, for whatever reason, it needs to be done at least 12, and preferably 24 hours before the meeting is scheduled to start.
(Response from Ivar) I think we agreed that I would help out by sending a reminder a day before the call with a link to the agenda. It is everybody’s responsibility to add to the agenda. And an empty agenda means we won’t do the call.
Sometimes it’s worthwhile to have a meeting even if there is no official agenda. Consider it an “unmeeting” where the agenda is formed at the start of the meeting or during the flow of the meeting. If it turns out to be a failure, we can stop doing it.
(Respone from Ivar) I agree! it is a good idea. I am happy to do so for future meetings.
If the meeting ends early, consider leaving the zoom running and displaying a banner that says the meeting ended early. If there’s confusion about the start time and the meeting starts later than expected for some people, consider starting the zoom conference early and display a banner that the meeting starts later, e.g., due to DST changes.
(Response from Ivar) The meeting is set up so it does not require a host. Anybody can log in to the meeting at any time. I won’t be able to attend the meeting every time due to travel (not currently, but usually…). So facilitating this would be up to someone in the group.