Minutes June 23 2010
Jeff Licquia (LF), Stew Benedict (LF), Alan Clark (Novell), Mats Wichmann (Intel)
- LSB 4.1 status report.
- The ongoing packaging discussion in lsb-discuss.
Jeff: 4.1 status page:
Robert wrote a couple of man pages.
Infrastructure work - some issues with building 4.0 bits of misc-test, where packaging bits are getting pulled from devel. Need some fixes to address this.
Jeff: Have several other projects going on that are slowing down some of the LSB work:
Meego, some work to make LSB compliance tools more generic, and another project. Hopefully will be able to get resources back on LSB soon.
Jeff: There's been some discussion on the packaging list about a new approach to the Berlin API: psys.
Mats: It's great that someone has come up with code. Is there anything we (LSB) should be doing? Should we even worry about it until distros adopt it? Or should we try to be somewhat supportive of it, since someone has done work to support something we proposed.
Jeff: A couple of issues: usual time and resources issue. Do we consider this as an open proposal, that we want to encourage to continue? Issues that were brought up at the time are still valid (gui installers working in conjunction with the native packaging system).
Jeff: Some responses on the list question why this is needed, why don't ISVs just make native packages?
Jeff: Some lack of participation/input from ISVs.
Mats: I think characterizing it as "something ISVs really want" is not quite correct. It's really more of a benefit to the system/administrator to keep the packaging database in sync.
Alan: Maybe we should invite the psys author to a call and engage him. See his plans, help guide him a little and maybe shop it around some for him.
Jeff: What do people think? Invite him to a call?
Mats: He knows the problem, that distro buy-in will take significant time. New approach deals with this to some degree. The other problem is whether ISVs really want this and will use it.
Alan: Instead of pulling him into a meeting, I'll try to ask him some questions via the mail list.
Jeff: Antyhing else?
Alan: Java. Do we believe we can fix the Java problem, or are we beating a dead horse? Can we get patches accepted into OpenJDK to build a compliant JVM?
Jeff: We made LSB building an option. If I understand Robert's issue, a "normal" build should just be a compliant build. I think the "ifdef LSB" was used to make things more palatable to the OpenJDK folks. This would get the code in, and hopefully with testing would eventually go away.
Alan: I think we need Robert to have this discussion. I may have interpreted his comments differently.