[Halld-online] Proposed production directory structure and build system for readout lists

[Alexander Somov]

Hi all,

There should be only one primary (and one secondary) readout list
(with some rare exceptions). Electronics board configurations will
be handled by configuration files, which are configuration specific.
We (both Sergeis, Dmitri R. and myself) are working on this
as well as on the interface with the run config DB. The location
will be discussed later.

Cheers,
        Alex



On Fri, 15 Nov 2013, Serguei Boiarinov wrote:

> One remark from Hall B experience: multiple copies of BCAL_rol.c etc rols were found confusing,
> at some point we decided to use one source code and ifdef flags set in makefile, so you can generate
> BCAL_rol_cosmic.so, BCAL_rol_prod.so etc from the same source. Usually 95% of those rols are generic.
>
> ----- Original Message -----
> From: Elliott Wolin <wolin at jlab.org>
> To: halld-online at jlab.org
> Sent: Fri, 15 Nov 2013 17:11:23 -0500 (EST)
> Subject: [Halld-online] Proposed production directory structure and build system for readout lists
>
> Hi,
>
> As the ROC's come online in the hall I realized the existing build
> system needed a slight modification to support multiple sets of readout
> lists, and the fact that each readout list must be converted
> individually to a shared object file (typically groups of files are put
> into a single shared object file).
>
> I modified the scons-based build system as required, tested it, checked
> it into the online repository and distributed it into our development
> build area.
>
> The proposed production directory structure for readout lists is shown
> below (design based on discussions with Mark D and Dave L). Readout
> lists reside in the production builds hierarchy as a standard package
> named "rol".  Readout lists are built with our standard online build
> system, but note that they must be built for the same operating system
> as is running on the ROC hardware (currently CentOS5).
>
> Below "cosmic", "fcalTest" and "prod" refer to sets of readout lists,
> and different sets can have lists with the same name (all names below
> are for illustration purposes, e.g. I don't know if we'll need a special
> list just for BCAL or FCAL).  Each set has its own SConscript file, as
> usual.
>
> Another proposed package, rocLib (or rocUtil or something like that),
> would hold libraries, headers and other code common to all readout
> lists, as well as utility programs meant to run on the ROC's.
>
>
> $ tree /gluex/builds/devel/packages/rol
>
> |-- doc
> |-- README
> |-- SConstruct
> `-- src
>     |-- cosmic
>     |   |-- BCAL_rol.c
>     |   |-- FCAL_rol.c
>     |   |-- SConscript
>     |   `-- TS_rol.c
>     |-- fcalTest
>     |   |-- BCAL_rol.c
>     |   |-- FCAL_rol.c
>     |   |-- SConscript
>     |   `-- TS_rol.c
>     `-- prod
>         |-- BCAL_rol.c
>         |-- FCAL_rol.c
>         |-- SConscript
>         `-- TS_rol.c
>
>
> Comments...?
>
> Thanks,
>
> -- 
>
> 				Sincerely,
> 					Elliott
>
>
> ================================================================================
>
>
>  Those raised in a morally relative or neutral environment will hold
> 		    no truths to be self-evident.
>
>
> Elliott Wolin
> Staff Physicist, Jefferson Lab
> 12000 Jefferson Ave
> Suite 8 MS 12A1
> Newport News, VA 23606
> 757-269-7365
>
> ================================================================================
>
> _______________________________________________
> Halld-online mailing list
> Halld-online at jlab.org
> https://mailman.jlab.org/mailman/listinfo/halld-online
>
> _______________________________________________
> Halld-online mailing list
> Halld-online at jlab.org
> https://mailman.jlab.org/mailman/listinfo/halld-online
>