Talks at EPICS meeting

Agenda items for V4 days (reminder, we have Monday and Friday). Who is going to be at
one or both of these.

4.4 progress other than type identification.


Attendees: AJ, GS, MK, MS, RL, DH, GC

Scribe: AJ


AJ: “EPICS Base 3.15 and Beyond

GS: Physics Applications talk

MS: Remote at V4 meetings only

MK: What does Bob want? Can give talks for MS

MS: Should be talk on pvAccess updates

GS: Anybody talking about NT & Type-ID?

MK: There should be.

AJ: Nothing explicit listed on Saclay website for offered talks. David Hickin?

MK: He will be there, there is a topic on Agenda for AD Services.

AJ: Assume there should be an update talk

AI on MK: Ask Bob what he wants.

DH: Talk on AreaDetector V4 Transport Service

AJ: Ralph any talks

RL: Nothing V4 related. Bob put a topic for pvaSrv, reluctant to repeat the same talk again.

GC: Will be there, talk on DIIRT, like the July V4 dev meeting. Talk on NTTypes, VTypes & CSStudio updates.

AJ: Greg said Model & Enterprise Data Services at SLAC

Murali has Archive and Directory Services at SLAC

Bob for V4 overview talk

V4 F2F Meeting Agenda Items

RL: PV Request Structure, & GW’s misunderstanding of who’s developing a Gateway

MK: What do you mean?

AJ: Expressing 3.15 server-side filters in pvRequest objects

RL: Another command-line language for expressing PV Request items?

RL: Since a pvRequest is a pvData object anything is allowed, but don’t want to allow just anything, need a standard.

AJ: Release process, adjustments in light of V4.4 release experience

MS: NT IDs. Can we release V4.4 without agreeing on these?

MK: That would bother me, might have to change in the future.

AJ: V4.4 release not done by then anyway.

MK: Proposal see my email on 9/28/2014

MS: Need pre-release ASAP, but don’t want major changes after 4.4.

RL: Waiting in the rain for 3.15 and V4.4…

GS: Issues with Diirt & CSS, waiting for the new release of pvAccess.

MS: Once the pre-release is out you can start work with that. Also working on formulas for NTndArray, show image in CSStudio.

V4.4 Progress

AJ: Are we ready for pre-1 release?

MS: Yes, need to change ID to current state (2014), then can release

MK: Why not make the changes is my msg, then release? Only Greg is against it

One URI for all the NTs




MS: What does GC think? Change all the type’s IDs whenever there is a new spec document, or only change the ID when the type definition changes?

GC: The latter.

MS: Greg says composed NTTypes should share a common version. This is nice, but...

DH: How about both versions have their own type, only increment

MS: Not many compound types. Don’t want to make design complex to solve small part of problem. Most use cases should be simple.

GC: When you add one type, you change everything and none of the old clients work any more, even though they don’t use the new type.

AJ: The advantage of EPICS is that the DBRtypes have been the same for 25 years.

MK: Can have more complex types and DBRtypes, evolutionary phase. Those changes shouldn’t affect the basic types.

GC: Even if the doc only changes every 4 years… doesn’t seem sensible.

MK: For pre-1 go with MRK’s proposal, discuss at the F2F meeting before final release.

AJ: Happy with MK’s proposal. Vote? Unanimous

Resolution: Follow MRK & MS’s proposal for URI, then discuss at F2F meeting before final V4.4 release.

AJ: Talked to SV, pvaPy is ready

GS: Interested in state of Python support for NTs. pvaPy has implementation of NTTable, should be compatible.

AJ: Ready for -pre1? Make these changes and release -pre1.

MK: End of today will have pvDataJava, pvDataCPP, normativeTypesCPP.

DH: Need to update Makefile and configuration script, add normativeTypesCPP

MK: Please do immediately

AJ: Aside, jenkins/cloudbees_build still including pvIOCCPP.

GC: Question about releasing EPICS v4 JARs in maven central?

MS: We should talk about it any time from next week.

AJ: Everyone knows what to do?

MK: Will update spreadsheet, then not sure what happens. Will be creating a release branch.

AJ: OK, fire.

Meeting adjourned at 10:56 Chicago time.