EPICS V4 TELECON 4-NOV-2014 MINUTES

===================================

AGENDA

1. Status of 4.4

 - pre1 cpp

 - items remaining for rc1

 - What shall mark when we build rc1?

2. Binary distribution for Win and Mac (Matej)

MINUTES

 

Present: GW, AJ, DH, MK, MS, GS, RL

Scribe: DH

Chair: GW

NEW TOPIC (Status of 4.4)

=======

GW: Have pre 1 Java C++. Build process runs OK.

Mistakes made in version numbering but too late now. Module numbers are not connected with distribution number.

MS: Created automatic Jenkins build. Windows and RTEMS tests pass.

AJ: Will attempt to run VXworks tests.

TOPIC DIGRETION: Binary distributions.

MS: Mac compiles from source. Incompatibilities in C++. Binary distribution: Programmers will build from source. Binaries for utilities for Mac and window. Not sure whether to create installer or just zip file.

GW: As we continuously build should be problematic to add to binary distributions for each new mac target

GW: Not worth holding up 4.4 for this - but nice to have.

GW: Need statement of what architectures supported. Spreadsheet. Statement.

MS: For Pre1 it is our list of jenkins builds.

AJ: Vxworks 5 not supported.

AI on MS (AJ). Make list of architectures compatible with 4.4.

MS; Need to release December 1st

BACK TO STATUS OF 4.4:

GW: Any coding want to do for RC1

MS: No

MK: I have 2 things

MK:

pvaSrv

   It crashes if given a bad pvRequest.

      RL: Intends to fix this (and other issues) by 4.4 release

   Is Busy record a problem?

      MK: asks DH if recommended work around works. DH: says it does not.

AJ: What ca call does it make?

MK: pvaSrv  needs an additional option to process but don’t wait for record processing to complete when a put is performed.

Note that pvaSrv does not use ca for puts but directly calls dbAccess.

MK: Can RL add this feature.

RL: Don’t want to add feature to already undocumented standard (pvRequest).

GW: Ralph is best person to define pvRequest

MK: Only pvRequest options being used are “process” and “queue size”, plus the list of fields required can be specified.

GW: Can what we have be written down.

MK: Tried this and every time people say that’s not what they want.

GW: Can RL write down a half page version of what he wants.

MS: For 4.4 just need to make up a name for “nowait”

MK: Only change required is to pvaSrv

AJ: Need to document the data structure as well as the parsing of this request.

MK: Already documented in pvRequest.

Resolution: Add new option for pvRequest for waiting for processing or not (wait). Default of wait = value of process . This will go into 4.4

rc1 = pre1 +

  1. new option in pvRequests as above
  2. bug fixes in pvaSrv, esp crash on bad pvRequest

   3.        Convert. Should copyUnion do shallow or deep copy?

MK: Variant union does deep copy. Regular union does shallow copy. Complex arrays are shallow copied.

DH: Regular union should be shallow copied in copyUnion

MK: Let’s make resolution in next couple of days.

AI: By Thursday evening make decision on copying unions.

GW: RC1 by 12th?

AJ: Can do RC1 on 18th to give us 2 weeks.

DH Pre2 in a week? Then RC1 18th

GW: Need code freeze at some point:

Resolution: Pre2 11th. RC1 18th

AJ: Do we need to announce pre/rc more widely.

Consensus is only for  RC 1 announcement.

AI: on GW, write to RL to ask that mods to pvaSrv are readied for pre2, so Dave can test prior to rc1.