============= Meeting 04-Apr-2012 =============


------------------------ Minutes ----------------------------

Present: DH, MK, MS, GW
Scribe: DH

********
NEW TOPIC: EasyPVA Review
********

MK: To get EasyPVA to build
command line hg clone
start eclipse
file,import easypva, Java part
hg/alphacpp/easypva

GW: Do I only need Java part to run tests

MK:Have to nake sure pvdata and pvaccess in projects.

GW: Have multiple methods of building  makefile for C++, Eclipse for Java, others?

DH: Maven?

GW: Didn't seen Maven in alphaCPP

GW: alphaJava in alphaCPP - love that

MK: How many repositoriories?: 
GW:Re look at locations later along.

MK: Do now or , do test, make release.do it.

MS; Do it after this reorg - else break something and have a mess.

GW: Finish reorg, test, make release.

MK: Use easyPVAVA is our test.

GW: You mean alphaCPP?

MK: Alpha is single space for all new ideas, rather than have multiple repos.

MK: Everything in alphaCpp/easyPVAA can be used as test for multiultiple languages

GW; Generalise toe to other modules will turn out to be complicated.

GW: Show me how to download ad pvIOC and build Java or CPP or both.

************
NEW TOPIC:  MoveField name and unsigned status
************

MK: Builds but not running correctly - problems with easyCopy. pvDataJava. pvvAccess java. pvIOC java builds but doesn't workrk.

MS:  Not sure about ppvServiceJava. On MS side side pvAccess works corrs correctly but doesn't communicate with cpp becaecause haven't touched it yet.

***********
NEW TOPIC: Eget and Normative Types Progress
************
MS: I'm ready to do eget. How should a normative type self-identify?
GW: Let's use the pattern in rdbService as an outline, then change it when we decide how
self-id will actually be done. see
https://github.com/epics-base/exampleJava/tree/master/src/rdbService/RdbClient.java

GW: For eget will it  it be able to get v3 and v4 pvs.

MS: In CPP. V4 only. 

GW: How will it be done?


GW: If I type a versionsion 3 PV how does eget establish whether to get data.

MS; Just do V4 a4 at moment. In future broadcast on both.

*********
NEW TOPIC: Plaform support
*********

GW: MS thinks shouldn't support Windows or VxWorks
Think need to help get VxWoks 6 to work.

MS: We can't say we support VxWorks because none one have us has compiled it.

GW: Can we say we don't run regression test on windows. We e believe VXworkss 6.? and window version ? will work with makefiles provide.

MS/GW: Can't say support it unless we regression test it.

GW: Are we agreed that will provide table with compatibility of each module and a clue how to do it (e.g. tr1/boost shared pointer)
Can MS providevide it.

MS: OK

***********
NEW TOPIC: Maturity matrix
************

GW: Need AI - probably to me. MK/MS to check out.

************
NEW TOPIC:  Performance Measurement
************

GW: Number of process gets goes uoes up with number of channels.

GW: Data isn't normalised - odd way of showing it.

GW; What hat about gather service graph speed of transaction (ms) goes down with channels.

MS: Will get all at once. Will happen in parallel up to some limit.

GW: Think divividing by number of channels.
Will have to ask Gubao.

************
NEW TOPIC: EPICS V 4 Meeting at SLAC
************

Few members going.

GW: May look to cancel eV4 meeting at SLAC epics meeting.
 Emergency move to Oxford? (3 days).
 
 DH: I'll look at possibility.

************
NEW TOPIC: EPICS V4 Test and Development Support 
************
GW: Can Jenkins support starting a V4 server and then making test acq from it?
MS: Yes, you just put the serbver startup in the junit test, and then make acq against it. 
GW: So we need a test server against which we can run regression tests.
MS: There is one in pvAccessCPP there already is a test server named "testserver"
GW: Does it automatically run when you do a mercurial release of pvAccessCPP?
MS: Yes, and the same is true of java.
GW: Does it do channelRPC?
MS: Yes, just a simple example though.

MS will write an email on how to start it standalone and as part of regression tests after a relase.

Meeting ends.