EPICS V4 Meeting, June 3rd, 2014
1. Review the planned features of EPICS V4.4 . That version has been modified
to emphasize features functionally (what a user can do, not what the sw does systematically),
and adds features not part of the proposed changes.
2. Review agenda for BNL meeting.
Link to the agenda has been added to the WG page .
PRESENT: AJ, DH, MK, MS, RL, GW, GS, NM, GC (joined at :35)
MK: To RL, will you do the changes to pvaSrv? MK & MS are ready to merge this friday.
RL: No, not going to be ready.
MK: OK, [mk] will do it.
TOPIC: dbGroup being in 4.4
GW: Will dbGroup be in 4.4?
RL: Not according to what we decided at at the last F2F meeting. Then last week we said someone would clone the existing dbGroup code and make it network-accessible.
MK: Would be put into pvDatabaseCPP.
GW: Functionality of dbGroup could be implemented by a pvaSrv running on the IOC.
RL: This is not multi-multi, which needs 3.15.
GW: Needs sync across lock-sets.
RL: Yes, existing code can do this
MK: We probably want to keep these as 2 separate things.
GW: pvDatabase is an esoteric thing, people might want a light-weight thing. Still want dbGroup without atomic operations across lock-sets.
GW: Will we have multi-multi across IOCs
MK: Multi-channel, we can do that, inside IOC or as stand-alone.
RL: Learning curve for pvDatabaseCPP is not as steep as GW things, it’s really just a library that this tool will use, so channel provider writers don’t have to reinvent the wheel.
RL: The existing multi-channel did special things with the IOC database if it saw local channels; we don’t think it’s a good idea to put imperfect first version into the code when the next version will be very different. Make this a separate service which maybe can run on the IOC but doesn’t have to be.
GW: Ok, as long as we have the functionality
RL: I can talk on dbGroup at the F2F meeting, and will talk about the API I’d like to propose.
GW: Ok, not on the first day though since it’s not in 4.4.
GW: MS please check my language.
MS: Protocol specification handles lot of code, encapsulated in a class (abstract), just need to add transport code (read/write) to this. Codec code was heavily tested in Java, ported this to C++.
GW: Where is Damien?
MS: Need to ask him.
Topic: Pluggable PVAccess Security API
GW: Will we have a plugin?
MS: Have to do something to test the code, maybe implement what CA has to begin with.
RL: Yes, this is what the first plugin needs to implement so we can use this on the IOC.
GW: Is this a stretch?
MS: Yes, but it will be put in.
Topic: Simplified eget tool
GW: I do currently get CA data by giving a URL, this has to be the fundamental way to interface with web tools etc in the future. Need to work out how to express REST in a V4 URL.
MS: Will be lots of pv-tools which are really for developers, not end-user tools. For physicists we will provide e-tools.
GW: As long as e-tools are complete; need put and RPC.
AJ: In favor of e being a higher-level program that just calls sub-tools.
GW: User said it’s useful to have one tool and everybody agrees on and is the same in all labs.
MS: Git has extensions, nice. Can add lab-specific commands etc.
GW: One too-set has to be able to do everything.
MS: Getting too much work.
GW: I need easypva multi-channel, will try and add that myself. MS priority is security.
Topic: NTNDArray, pvDatabase, pvPy
GW: NTNDarray & pvDatabase will be in 4.4, pvaPy
AJ: pvaPy would need Python development libraries/headers.
GW: Can we include pvaPy but have it not kill the build if those are not present.
MK: Internal discussion in July for what to do about pvRequest.
MS: Smart gateways, etc.
MK: We need a serious discussion before we decide, this won’t be in 4.4 (not perfection).
2. Review agenda for BNL meeting.
GW: GS, will you be giving a talk?
GS: Need to talk to Bob what he wants.
GW: I’ll do that, prepare a developers talk and a users talk of CSS. Main one is Users.
GS: Ok, will include: How you have a pure python script running and have it update CSS, integration of services etc. All depends on where I’m directed to work.
GW: I’ll add that talk then, let me know if you can’t do it. We do have a lot of stuff to cover though.
Meeting closed at 09:53 PST.