STAGE 2001-9-29 meeting summary


Hermes design:

James has some objections to it's use of a custom protocol, and it should rather use Atlas. This would make it alot easier to add support for hermes in clients.

According to bryce, Hermes should be able to be compiled out of STAGE via a compilation flag, and still be possible to fully operate a game system.

A volunteer is needed for Hermes, else it will just sit and rot. But this should not have any effect on the rest of the STAGE development.

New STAGE coordinator:

As Bryce is not activly developing STAGE currently due to other projects, Morgenes had been coordinator for a while. But as he's been out of the loop for many weeks now, it was decided that a new coordinator was needed.

Alas, James is now our new chief fool.

Goals for the next few weeks:

James has a list of goals he'd like to share with us, but he claims he can't do that untill he get's a real request tracker like Bugzilla. Bryce will be integrating RT with Eidetic in the coming months.

The game 'Pong' in the battleplan will be replaced with 'rim-hack' since the requirements match better with what can be acomplised in the near future. Simulating a Pong game on STAGE requires work that will not be finished any time soon.


Bryce mentioned that OSDL is using standard benchmark tests and other sorts of "workhorse tests", that can be set up to run and produce reports and graphs.

Use of the macro ASSERT in the code:

A couple of places a macro called ASSERT is used in the code. The issues with this is that when failing, it calls exit(1), which leaves no stack for gdb to debug. Soon after the meeting, the macro was rewritten to use assert instead.

Bryce mentioned that they're implementing database tables for Circe in Eidetic and that it's his intendt to ensure that it works with STAGE. And he was curious about what state STAGE's database currently are in.

Summary by Rakshasa