EPICS V4 Telecon 27 Jan 2015

----------------------------------------

Agenda

---------

shared_vector (Michael)

Gitmo

new CSS V4 integration (Dave, Matej)

Java Bundle (thanks Ralph)

Present: MS, DH, RL, GC, GW

Scribe: DH

Chair: GW

Minutes

----------

*****

NEW TOPIC: Shared Vector

*****

GW: MDs waiting for feedback from SNS

MS: What MDs has done is useful and recommends  to include in pvData.

GW: Make decision through lists. I’ll send the mail;.

*****

NEW TOPIC: Gitmo

*****

GW: Moves to git seems to be what’s being developed by group. But don’t want to move until have found out what plans for base are. Don’t want to move unless there are descriptions of what to type.

GW: MDs scripting move. Not necessarily doing documentation.

RL: MD proposing changing development. Having dev and release  branches in same repos  bad idea.

If into using GitHub works best if use a pull request model using web interface. Hg flow models keeps different developments  in one repo on different branches. GitHub model uses private repos.

GW: Need to have a plan before move.

GW: Pull request model works well when not sure of quality of commits. (e.g. in open source project).

DH: Not against move, but this is not high on my list of priorities.

RL: Should wait for base.

DH: +1

GW: Propose  1) make plane before branching.

2) Wait for base. Or at least wait for input from base.

RESOLUTION: The WG shall wait for input from EPICS Base on Base’s plan for VCS before deciding on a Git (and separately GitHub) move.

Passed.

RESOLUTION: The WG shall prepare a statement and help on merging (merge model) and release before making the move to Git.

Passed.

GC: I could write down a merging model [as used by CSS] and email to list.

*****

NEW TOPIC: new CSS V4 integration (Dave, Matej)

*****

DH: CSS for V4 worked. Able to get image from NTNDArray. Able to use formula for …

DH: One thing that didn’t work was getting an NDArray out of an NTNDArray.

DH: Also, image displays didn’t have scaling, which means you need an enormous widget - bigger than the screen - in order to display it.

GC: Ok, will look at that.

DH: Is very happy with what he has. Most of what DH needs is working.

DH: Also putting to an enum pv by index rather than string didn’t work.

GC: Note that in CA, putting to a string has different behaviour depending on whether the pv is local or remote. GC doesn’t know behaviour of pvAccess.

MS: Possible I didn’t implement putting to enum pv by index other than string.

GC: Iter want internationalisation:  Date format. Decimals.

GC: Having everything client side is probably best bet.

*****

NEW TOPIC: Java Bundle

*****

GW: Ralph’s gone, but thanks. Haven’t have time to test but looking forward to testing for next release.

*****

NEW TOPIC: Next release

******

GW: What goes in.

MS; Fix to C++ code that allows SNS neutron server to  run without crashing. May not need a release because SNS are happy to work off latest repos.

MS: Suggest do a new minor package release soon.

GW: Do release in next week or so.

DH: Pre 1?

MS: Not necessarily.

GW: Go straight to RC1.